Using bad numbers

From Sunny.


Nope, wrong.

That is a display of market incomes. It\’s from ASHE, compiled by the ONS. Perfectly good numbers, but it simply isn\’t incomes. It is pre-tax, pre-benefit, market incomes only. As I say, within those confines, they\’re perfectly good numbers.

But they\’re just not \”incomes for Britons\” because we have a tax and benefit system which redistributes those market incomes.

Is this more than just pedantry? Yes, actually, it is. I think most lefties would argue that we should have higher taxes on the rich and more money going to the poor. That we should have more redistribution a la Sweden rather than no redistribution at all like Hong Kong.

Say we moved to either of those extremes from what we currently have. How would the graph change?

Not one single iota.

So it\’s not actually a very useful graph, is it?


35 thoughts on “Using bad numbers”

  1. Do the useful versions of those numbers exist? It’d be nice to produce a graph from them to compare.

    Tim adds: Yes, they do, but I’m not entirely sure where. Household incomes, post tax and post benefit, are certainly calculated by quintile and perhaps by decile. Umm, actually, there was a big report on it maybe a year ago. ONS or perhaps statistics agency?

  2. It’s a very useful graph for producing foam-flecked invective against the wealth-producers if you are a member of the Looting Party.

  3. It also ignores the composition of each income bracket. There might be lots of churn.

    The top 1 per cent bracket might be composed of many different individuals over a 35 year period.

    The bottom 50 will surely have lots of churn. Recent graduates command low salaries, but then enjoy rapid income increases.

    This chart implies (to the unwary) that there are people called the “1 per cent” and others tattoed with the “bottom 10 per cent”. Sunny falls for this in claiming chart this tells us something about “most Britons”. In fact most Britons will have traversed at least two of these income brackets during that time. Some will have been in all four segments at various times.

    Claiming income charts tell us about individuals is no more valid than dividing cars in to ones which do 30mph all the time and ones which move forever at 70mph.

  4. If my memory is correct, recent US figures showed that of the top 1% in 1996, 43% of them were in the top 1% in 2005.
    The figure for the bottom 20% was similar; around half had moved out of the band in 10 years.
    I can’t see why the UK would be hugely different.

  5. .. and the ‘bottom 50% line’ has risen significantly, somewhat disproving the ‘incomes for most Britons have stayed the same’ text. Still, never let the data get in the way of class warfare.

  6. According to this graph London Underground train drivers are in the top 1%. I know they’re on a good screw but can this be correct? If so, will the Occupiers be fussed about them?

  7. Churn alone makes the graph useless. Student households make up many of the “poorest” households by income, because they are studying and not earning full-time, but as a statistic that’s pretty much meaningless. When the IFS did their calculations re the new student finance farrago, they sensibly crunched their numbers based on projected lifetime earnings, which eliminates the churn problem but replaces it with a “how do we project – and assess the probabilities – of individual earnings trajectories” problem.

    Bottom line is that it’s actually quite difficult to assess who “the poor” are. Assessing who “the rich” are is easier of course, but if we make the natural switch from measuring income to measuring wealth, we end up conflating wealth (especially value of pension fund!) and age. Even if we built a new society, reset indidual wealth to zero, and then imposed permanent perfect income equality, in a century’s time we’d still end up with huge wealth inequality – some people save/invest more than others, and older people will have had more time to fill their pockets.

    Bottom line: (in)equality is multifaceted, can be defined in different ways, no single graph can really represent what’s going on, and individual trajectories (mobility) matter as much as snapshot aggregates.

  8. The Apiarist, Hugo,

    The source data for that chart is the ONS’s Annual Survey of Hours and Earnings (ASHE). This is seriously flawed because it only looks at PAYE earnings. It completely ignores dividends, interest, capital gains, etc. Many (most?) high earners structure their earnings to be paid as dividends or capital gains, not as income, for tax reasons. Thus they don’t show up in the above graph.

    Depending on your point of view this is either legitimate tax avoidance, or yet another example of how the 1% don’t play by the same rules as the rest of us.

  9. From a quick squint:

    Poorest decile:
    pre-tax, pre-benefit income: £3,888
    after direct taxes and cash benefits: £8,162
    after all taxes and all benefits: £12,816

    Richest decile:
    pre-tax, pre-benefit income: £100,155
    after direct taxes and cash benefits: £76,089
    after all taxes and all benefits: £72,934

    So the multiple from top to bottom is:
    pre-tax, pre-benefit income: 25 times
    after direct taxes and cash benefits: 9.3 times
    after all taxes and all benefits: 5.7 times

    25 times pre tax and benefits, down to 5.7 times after tax and benefits – that’s a hell of a lot of redistribution.

    Alternatively, barely worth working hard if you end up only 5 times as well off as someone who doesn’t work.

  10. Looking at it, I suspect the bottom end are even better off than the ONS say.

    They’ve got figures of around £1,000 p.a. per household for housing benefit and around £300 p.a. for council tax benefit. I’d expect both of those to be much higher.

    If I can find the figures, it would be worth comparing the total benefit from the ONS figures against the total cost of the benefit provision.

  11. Also the ONS figures are flawed, because apparently half the people in the bottom 10% have negative incomes – i.e. business owners who made losses.

    These are only temporarily in the bottom 10%. I’d say you should be using an average income over a few years for them, so they’d come out of the bottom 10% (raising the average there) and go into higher bands, lowering the average there.

    That would be more accurate reflection of the true income for the self-employed.

  12. Aha, here we are.

    Total cash benefits received, according to the ONS (assuming my sums are correct) – £146 billion.

    Total cash benefits paid out (2010 Budget) – £186.6 billion.

  13. The reality is that if you are fortunate enough to come from the benefit aristocracy you can inherit what is basically free accomodation in central London, something that would cost a non aristo, say, £1000pcm. That’s 12k a year, on current taxes you need c£20k of earnings just to pay that. Plus you don’t pay council tax or any of the other wonders of modern Britain, even getting your heat and light thrown in for free in many cases. Work on a building site for around £100 a day cash and you take home £25k a year, which again at our wonderful tax rates leaves us with the ludicrous situation that the unskiled labourer on £100 a day cash in hand has the same disposable income after tax and accomodation as the aspiring, hard working tax paying memeber of society who used to think they were doing OK on £60-70Ka year!

  14. Have a salary of over £45K and join the 1%. Something’s not right there. That would mean that all the NHS managers and all the council big wigs are in the top 1%.

  15. SadButMadLad, I think you’re misreading the graph (mind you, the graph is possibly made to be misread).

    I think the top of each coloured section is some sort of income level for that section.

    It looks like this should actually be a line graph, with four coloured lines; filling them in is confusing, but makes it look like the 1% have a huge chunk of the graph.

  16. SadButMadLad: Yep, something is not right with this graph.

    I’m not even sure what it is purporting to show. The starting point for each income bracket? If so, why does the top 1% have a maximum?!

    If you earn £500k you are off the chart – beyond the max of the top 1%.

    So are these averages?

    Sunny has provided an incorrect link to the original source. The correct link to the organisation responsible is:

  17. This I think is the first time ever I have disagreed with you.

    That graph indicates a trend. If redistribution offsets that then it is useful as an indication of redistributive policies.

    But surely if one argues that the balance of the economy is shifting in an unfavourable direction this is supportive. If one thinks it should be organised in a way where redistributive policies are not essential then it is evidence of that failure.

    It is therefore a graph supportive of government intervention. At least if greater equality of earnings is what you desire.

    It proves the point many make.

  18. The more I look at this data the more obviously rubbish it is.

    According to HMRC, the top 1% earns in excess of £150k pa (the Top 10% earn more than c.£40k pa).

    So how can the graph say the top 1% earn £100k (at 2010 prices) when HMRC tell us the answer is actually over £150k?

    re the distribution point, the HMRC data also tells us that the share of total income claimed by the top 1% has rocketed from 9% in 1999/2000 to…erm…9% in 2010/11.

  19. @Richard, you’re right. I’ve misread the graph – probably because that is the point of the badly drawn graph – to make the top 1% look massive because of all the colour. It should be a line graph not an area graph.

  20. Worstall, you are truly an irrevelant-wad

    How many times have you used a load of balls to try and pretend you’re clever?

    Knowing how liberable you are, and the rest of the [email protected] that seem to buzz about your regions, of the fucking reality, I’m truly astounded at the really, rilly, idiocrasy of your outcome, both pedantic, and intelligence-wise.

    Do any of you live? Are you alive? Do you not have senses?

    Or something other than blind bollocking bollocks?

    Jesus. oh ha ha Jeebus. oh ha fucking ha.

  21. Arnald, you’re a fascinating chap. I’d be very interested to hear your life story, what you do for a living etc…

  22. of course, for my post to make any sense, the word I should have used is output rather than outcome. I fear I may have been overwhelmed by the sheer weight of pretence on here.

  23. The chart is badly drawn (I think the 1% is above the top line, and the 10% above the second line etc) but such data is clearly informative, it shows how the market economy is performing in distributional terms.

    It’s somewhat bizarre to think that you should only show data that is post-government intervention.

  24. Matthew2,
    Why would we*not* want to know the impact of our actions i.e. show the post-government intervention numbers?

    Surely the lived experience is what matters to people, and that is a post tax position?

  25. @Matthew2: So if the graph showing pre-distribution indicated incomes were growing at the bottom and falling or flatlining at the top, and the post-distribution graph reversed all that and meant the poor were worse off after tax & benefits and the rich better off, then that would be OK would it?

  26. Pingback: FCAblog » Eoinomics: committing Worstall’s fallacy

Leave a Reply

Your email address will not be published. Required fields are marked *