Monday, 30 May 2011


One of the key assumptions in the dominant narrative of the Greek crisis is that poverty has exploded to unprecedented levels.Frankly I never expected groundbreaking new levels of poverty because our social welfare system is one of the most inefficient (in terms of poverty alleviated per Euro spent) in the developed world and thus much of the money lost is money that would have been diverted to better-off recipients in the form of either welfare payments of wages. More on this here. But there can't not have been a substantial rise in poverty, and any rise is bad news.

As it happens, a new study is out that models the effects of policy changes on personal incomes from 2009 to 2010. Now if this was a country with a proper policy industry, this study would be cited left, right and centre, and its findings would be spun by all sides in whichever way. But there is no such thing in Greece, so I guess I have to spin by myself.

First things first:

The most common definition of relative poverty (also known as the standard poverty rate) is the % of persons earning less than 60% of the median (not average) income. This is relative in the sense that the poverty line is higher in more affluent societies. For a discussion of absolute poverty in Greece, you may want to take a look here.

By that measure, the study says, 20.1% of Greeks were living under the poverty line in 2009 and 20.9% were doing so in 2010. That’s a very small rise in poverty, in the order of 4%. However, this is not the end of the story. The authors suggest (and I think they are right in doing so) that a better measure of the change in poverty rates would be to benchmark against the 2009 median income, adjusted for inflation. This conveys much better the subjective experience of sliding into poverty – essentially we’re crudely counting the net number of people who have woken up to find themselves unable to afford some pretty basic things that they could afford a year ago. 

Benchmarked against this figure, those that are ‘poor by 2009 standards’ made up 25.2% of the population, which means a 25% rise in relative poverty. That’s pretty substantial.  The difference in the rate at which poverty levels rose under the two specifications comes from the fact that ALL of Greece saw its income fall in absolute terms in 2010.

So far, so predictable, although it’s important not to trivialise the issue of a 25% rise in poverty. I know some people have made it sound like this figure would be 2000000000%, but the real figure is still horrible news. But there are some findings that really are shocking.

Shocker no 1: Inequality actually fell in 2010, although this was largely because the upper middle class lost ground to the lower middle class (such as it is).

Yup. Although depending on what measure of inequality one chooses, the results can vary a lot. The top quintile of earners (i.e. those earning more than 80% of the population does; though check below for a hint at how well of they are) put even more distance between themselves and the bottom quintile (those earning more than 20% of the population does). But the overall Gini coefficient, the most commonly used measure of income inequality, actually fell. Can the same country become both more and less equal? Of course it can.  See how in the graph below:

Shocker no. 2: Public sector cuts reduced inequality

Yup. Why is this? Because 74% of public sector employees belong to the 30% highest-income Greeks. This means the civil service has almost the same share of marginally comfortable staff (realistically this is people earning a net EUR1,200 - 1,300 per month, see below) as the banking sector, the poster-boys of excessive compensation. See below:

Changes to personal income taxation also reduced inequality somewhat, while VAT was hugely regressive and cuts to pensions mildly so.

Now I know there's a grimoire of methodological caveats attached to all of these figures but to me they suggest one thing. It is easy to simplify the facts of the crisis to fit ideological frameworks; only if we dare to look the data in the eye, however, can we evaluate policy. I would add to this the suggestion that raising taxes is very bad for equality, on top of being very bad for growth. The best way, surely, to keep poverty at manageable levels ahead of the inevitable default, is to shift more of the emphasis of our adjustment programme away from tax and onto cutting spending. 


I know I said there's 'a grimoire of methodological caveats' to the above figures, but following a well-placed question from @side_shore I think I should point out the most important one.

Basically it concerns the accuracy of income reporting. The incomes used in estimating the above figures are of course themselves estimates. There is no database of Greek incomes after all. Basically, the authors have used EUROMOD to extract incomes as stated to the tax authorities. Assuming massive amounts of tax evasion and avoidance, they have adjusted these figures. The assumption is that incomes are under-reported by 1% for salaried employees, 0% for public sector pensioners, 55% for farmers and 25% for the self-employed (see pg. 7 here). These figures are not arbitrary but rather they are taken from this study.

The problem is that, as I mentioned, there is no such thing as a database of personal incomes, nor could there ever be. So what the authors of this last study did was compare incomes reported to the pan-European EU-SILC study with real tax returns in Greece and arrived at the above figures for under-reporting of incomes. The problem, however, is that while people are more honest when responding to survey questions than they are when filing tax returns, that's a far cry from being totally honest. If you make a buzzilion per year, you don't really want to tell somebody over the phone. In fact, if you earn that much you're unlikely to stick around to take the survey. Similarly, if you're very poor you may not want to discuss this over the phone, or you may omit benefits you are receiving from the state either because you don't consider them to be actual income or because you don't know what the particular benefit is called.

The authors of the first study I mentioned do recognise these limitations but argue that the incomes recorded by EU-SILC have been shown to match up to national aggregates. For a full discussion of the extent and sources of under-reporting, see this paper. It tests the data in places like Hungary and Italy, where people can be as crafty as the Greeks sometimes.

Finally, there's a big problem with the tax evasion and avoidance statistics because they reflect 2004 incomes. In 2004, the tax avoidance and evasion landscape didn't look anything like it did in 2010. The authors of the study I discuss here claim that avoidance got worse between 2004-9 and fell in 2010 under pressure from tighter inspection, but there's actually no way of knowing. In fact, it makes sense that the opposite happened as in 2010 incomes got tighter and tax morale (the crucial feeling that one is not an idiot for paying their fair share of tax, a crucial determinant of tax avoidance/evasion) must have fallen as the government's legitimacy fell rapidly. For more on the determinants of tax morale, see my past post here.

What effect these shortcomings would have on the validity of the findings (not the data, mind you) depends on the extent to which people in the top 10% and 30% of earners under-reported their income to EU-SILC. Despite being a European survey, EU-SILC is administered by ELSTAT and, upon being recruited to respond to it, one is informed that refusing to take the survey is punishable by law, much like dodging a Census. Despite assurances that all answers are confidential, this will have put some people's backs up. You can see the precise questionnaire for all past versions of EU-SILC here.

Bear in mind: the top 10% of Greeks are not a very exclusive club. To join them in 2006, you had to have a net income of about EUR36,000 per year (source). Let's call it 40k net in 2010 money. That's a good salary by Greek standards but it's hardly...

To join the top 30%, you needed a net income of about half that, EUR18,000. That's not a princely salary, and in fact I'm not surprised most civil servants get paid more than this; it's just not that much money. This of course also means that most people in the top 10% and 30% brackets are salaried persons who can't really hide their income from the taxman and wouldn't bother to hide from EU-SILC. That said, most of the income of this 10% is concentrated in a few households so if they are massive tax evaders or avoiders then the data is problematic. I don't know how bad the problem is but do look at the validation study I link to above; it should give you a hint.

So, yeah, lots of caveats.

1 comment:

  1. Keep up the good work man, it's a pleasure to read your posts! Only problem is that, for a late comer to your blog, orientating in it (given the titles you choose) can be a bit tough... But anyhow: Good job! :)


Please remember that I am not notified of any comments and will not respond via comments.

Try to keep your criticism constructive and if you don't like something, do tell me how to fix it. If I use any of your suggestions, you will be duly credited.

Although I'm happy to entertain criticism of myself in the comments section, I will not tolerate hate speech. You will be given a written warning and after that I will delete further offending comments.

I will also delete any comments that are clearly randomly generated by third parties for their own promotion.

Occasionally, your comments may land in the spam box, which may cause them to appear with a slight delay as I have to approve them myself.

Thanks in advance for your kind words... and your trolling, if you are so inclined.