Micheál Collins of the Nevin Institute is out with a new paper looking at the burden of taxation by income decile by tax-type, and the results are very interesting. From the piece:

Using data from the most recent Household Budget Survey, this paper estimates both the direct and indirect taxation contributions of households. The paper examines, individually and collectively, the direct and indirect tax paid by households across the income deciles, alongside the overall average household contributions. The data is presented at the households and equivalised adult level.

This chart summarises the findings nicely.

Update: Micheál has responded to many of the main points raised in the thread here.

By Stephen Kinsella

Senior Lecturer in Economics at the University of Limerick.

51 replies on “Ireland”

Interesting paper but pity it isn’t a bit more fine-grained about the top decile. There’s reason to believe that the low tax take from the 1% may be just as scandalous as the very high tax take from the lowest decile.

This is a good piece of work (I rarely use superlatives 😀 ) and an example of innovative thinking in research.

In the US, people like Mitt Romney, the Republican presidential candidate in 2012, used the statistic that 47% of Americans pay no federal tax, to dismiss them as “moochers,” when they pay several other taxes.

It’s always good to get the facts, whether people like them or not.

@ MH

There is also, of course, another way of looking at it i.e. by defining what is meant by income. One must assume that it includes, and can be made up entirely of, social transfers which, as you have pointed out on many occasions, are remarkably widespread, and generous relative to other countries.

In Ireland, one in five households suffers from “low work intensity” i.e. no one in gainful employment.

That VAT takes a higher share of low incomes than higher ones is not news.

I agree that the paper is a very valuable contribution but I cannot see where it addresses the issue I outline above. Although it does touch on impact of universal benefits in terms of the gain to those who actually do pay taxes other than VAT.

Readers of this report should be alerted to two very important features that call into question the ‘takeaway’ finding that the overall tax system is ‘regressive’.
1. The lowest household income group declared an average annual gross income of €9,887. This is less than the ‘income’ received by people on the standard rates of social welfare. For example, the individual rate of the non-contributory OAP is €11,388 a year. Who are the households that receive less than this? We need to know more about these households, on whom so much of the commentary provoked by this study has focused.
2. The study focuses almost exclusively study on taxes paid on income plus expenditure relative to income. Nowhere is the tax burden relative to expenditure shown. (On page 13 there is a footnote reference to an Appendix table that relates indirect taxes to expenditure, which it is admitted ‘mitigates to some extent the regressivity reported above’.)
If the author had included a table showing total taxes paid by households relative to their declared expenditure, a different picture would have emerged than the one that has received so much publicity. Households at the lowest end of the income distribution report expenditure that is significantly higher than their reported income. The bottom decile, for example, report expenditure of €18,495 – 87% higher than their reported income.
When total taxes paid are shown as a proportion of reported expenditure the picture that emerges is of a steadily progressive tax system. Total taxes paid by the the bottom decile came to 14.8% of their expenditure, compared with 50.6% in the top decile. The incidence of the total tax burden relative to expenditure rises steadily after the first two deciles and suggests a radically different view of the progressivity of the the tax system than presented in the study.

One thing that I think Micheal could address is the VAT gap Seamus mentions in his December blog post, which obviously throws off some of the results, and perhaps addresses Peter’s second point on reported expenditure somewhat.


“In Ireland, one in five households suffers from “low work intensity” i.e. no one in gainful employment. ”

Does that figure include a single pensioner or a couple of pensioners?
If so expect it to grow.

Neri said last year that 20.7% of Irish workers were classified by the EU as low paid (with the cut-off point being €12.20 per hour).

They also generally work fewer hours per week, although they wish to work more.

Two-thirds of private sector workers are in indigenous non-exporting firms while 56% work for indigenous non-exporting SMEs.


“The lowest household income group declared an average annual gross income of €9,887. This is less than the ‘income’ received by people on the standard rates of social welfare. For example, the individual rate of the non-contributory OAP is €11,388 a year.”

Peter as you well know the non contributory pension is well in excess of of the JSB or JSA for people between 26- and 65. €188 per week would put any single person beween 26-65 living on their own below the average you quoted
We know from the silc data that many unemployed people do live on their own (in bedsits)
Then there is also a small cohort of people that would be bringing the average for the bottom decile way down. People between 18-25 on JSA or JSB living on their own. Their annual income starts at €5,200 rising to the dizzy heights of €7,488

“When total taxes paid are shown as a proportion of reported expenditure the picture that emerges is of a steadily progressive tax system. Total taxes paid by the the bottom decile came to 14.8% of their expenditure, compared with 50.6% in the top decile.”
If the method of measurement you are using shows that the top decile are paying 50.6% effective rate of tax I would suggest that you method of calculation is wrong.
The Marginal rate is 52 and anyone paying anywhere close to that needs to hire an accountant.
According to the Revenue the effective rate from the top decile is in the mid 30’s.

Ernie balls request that the top decile needs to be deciled would also be a welcome chart. I think his instinct that many of the very very rich are paying less than the very rich may well prove accurate.

re income data:

The income data used is not actual household income but equivalised income, which adjusts for the size of the household: adults have a weight of 1.0. children over 14 0.5 and children under 14 0.3.

A few general observations on this:

One of the key issues here (and discussed by Seamus Coffey in his blog from last year) is what is the appropriate measure of household resources upon which to base these figures, income or expenditure. My understanding is that ideally some form of permanent income measure would be used (which we don’t have). Failing that, expenditure is preferable to income as expenditure is more likely to be related to permanent income than current income is. Plus, the HBS is really a survey designed to investigate expenditure patterns and I think most researchers would regard its expenditure data as more reliable than its income data.

The NERI provides figures for equivalised as well as non-equivalised income. When using equivalised income, I think it is also appropriate to weight by household size (which I don’t think the NERI analysis does). If you don’t, then you are regarding a household of one person to be as relevant to the analysis as a household of ten. Not sure what difference that would make to the analysis – perhaps not an awful lot, but I think it is probably better practice.

Regarding tax incidence, it is very difficult to pin down. Theory tells us that for a perfectly competitive industry it is 100% on the consumer, and so the approach in the Nevin paper is correct. When we move away from perfect competition, pretty much anything goes, though I think the balance of evidence is that incidence will be less than 100%, just about. The Mirrlees Commission in the UK thought that 100% incidence was a reasonable working assumption.

Finally, I wonder how much of the regressivity of indirect taxes arises from the socioeconomic gradient of smoking. Quite a lot I would guess, given that excise taxes seem to be more concentrated amongst lower income groups than VAT. Tobacco is probably the most highly taxes commodity (and there are strong efficiency and externality arguments for that being the case), but it is very regressive.

One of the issues mentioned by both Seamus and Peter is very important. There is a massive gap between reported income and expenditure at the bottom end of the earnings spectrum. Households in the first decile have expenditure around 80% higher than their disposable income, a difference of over €8,000 annualised.

I can’t insert a graph into the comment but the figures below are tax as a % of household expenditure (allowing for the assumption that the incidence of indirect taxes to fall completely on the household in the first instance with all figures equivalised by household size).

Decile Tax % of exp
1 14.9
2 14.9
3 16.2
4 16.9
5 19.4
6 24.0
7 27.3
8 30.6
9 36.4
10 50.6

This shows a very progressive tax system with households in the top income decile with total household income above 106,000 a year (or equivalised household income of just over €29,000) paying the equivalent of 50% of their expenditure in taxation.

It’s hard to tell from the data on first looks what the issue is here although it is common in cross-sectional household surveys. Given that there are no huge difference in household makeup the data (and particularly no case of higher representation of pensioners – or other household who can spend from savings rather than income – in the lower deciles) would seem to show some support the permenant income hypothesis.

The fact the study annualises data collected over a two week period means people in transition between jobs or who have just lost their jobs will be missclassified into low income deciles while maintaining their expenditure from savings or borrowing (It is clear households in the bottom income decile are funding expenditure 180% of their income from somewhere). The magnitude of this effect on the data would be compounded by the massive labour market turnover during the period of the survey (Q3 2009 – Q3 2010). Unfortunetely the survey has no data on household wealth or saving.

There is a similiar compositional issues in a recent TASC study on tax which included among its analysis of 2.1mn ‘taxpayers’ 800,00 people who paid no tax after credits (mostly occupational pensioners, social welfare reciepients and people who worked for very short periods during the year). That same study if it has looked at PAYE tax payers data and different family groups seperately would have come to radically different conclusions.

These temporal issues are likely the reason for the gap in VAT Seamus finds ie people eventually adjust down their income expectations over time and income expectations are likely to have changed radically over the period due to labour market transitions.

@ eamonn moran


Ireland sticks out like the proverbial sore thumb and there is no real mystery as to why. Decisions on taxation and social transfers are largely taken on the basis of political desiderata. PR and the STV ensures that the system remains firmly embedded and possibly impossible to shift, short of a change in the electoral system which would allow for consideration of tax and expenditure decisions on a national rather than a sectoral and local basis. The situation with regard to LWI has, of course, been made much worse by the economic bust.

The oddity is that the result of this haphazard process is pretty good in terms of overall equity and the progressivity of the overall taxation system. The real questions are (a) can we any longer afford it and (b) how can it be changed to permit it to continue in some form?

cf Chris Johns in today’s IT.

P.S. A lot of what he says seems to make sense to me but he is addressing his criticism to the wrong target; the major changes that are required remain a national responsibility. How long this game of pass the parcel can continue is an open question.


OK Before we cancel democracy ( I wouldn’t call first past the post democracy in a fit) lets take a look at the possible reasons that Ireland does stick out like a sore thumb in this.

Why is Ireland twice as bad as Spain and Greece whose level of unemployment is much, much higher?

Is it that people on lower incomes in Ireland are more likely to provide lots of offspring due to generous child benefit but in other countries with less social transfers they produce less than average as welfare is low?
Given the rate of population replacement is plummeting in many of these countries is that a model we want to follow?
Could this be related to the policy in this country of forcing women to proceed with unwanted pregnancies only on lower economic parts of society?

If that’s true are we perpetuating an ever larger underclass that are disaffected with society?

Really interesting Graph though.

@ eamonn moran: “lets take a look at the possible reasons that Ireland does stick out like a sore thumb in this.”

Try, Monica Prasad: ‘ The Politics of Free Markets: the Rise of neo-liberal Policies in UK, FR, D and USA. Quite Interesting comparisons and contrasts. The taxation contrasts especially! Maybe something for us in there? Or check out Seán Ó Riain: ‘The Rise and Fall of Ireland’s Celtic Tiger: Liberalism, Boom and Bust’ – particularly chapt. 5. These two about cover the ‘reasons’. And they are not exactly the ‘usual suspects’.

Looking at Table 2 of the working paper, expenditure exceeds gross income for the bottom 4 deciles. For the lowest decile the excess is 86% and for the second lowest 27%. The authors note as reasons “pensioners who may also be living on past savings, the temporarily
unemployed and students”. While negative transitory income may have some role in all of this as far as the unemployed and students are concerned, such large excesses of expenditure over income by 20% of households seem to imply that really large numbers in the population are temporarily poor from time to time, but recover relatively quickly. Either that or the income-poor have very large assets which they can draw down to sustain expenditure at 186% of income. So how this expenditure is financed is a bit of a mystery, especially as a lot of unemployment in Ireland is long-term

So the main conclusion of the paper – the central role of regressive indirect taxation – is largely driven by the “excess” expenditures of the bottom 20%. Is income underestimated for these groups? If for example we had data for pensioners’ assets as well as income we could look at the annuitised income from these assets and add this to income as conventionally defined.

Finally, are there reasons why people might under-report income? Maybe there are, and not just at the very top.

@John Sheehan

‘Finally, are there reasons why people might under-report income? Maybe there are, and not just at the very top..’

Black market perhaps ? Some years ago Frank Brennan the tax advisor suggested in a paper that the Black Market economy in Ireland could be close to 20% of all activity. They haven’t gone away you know !

It’s worth bearing in mind that the Household Budget Survey from which these data are drawn struggles to achieve a reasonable response rate. Despite the payment of a modest reward to those who complete the questionnaire and diary, the final response rate has declined from 54% in 1999-2000 to a mere 40% in 2009-2010.

Peter Stapleton, the use of sampling weights should take address low response, presuming such low response is not random. Actually I didn’t check, but I presume sampling weights were used in the paper.


I think that the answer is yes. The author advanced on RTE radio the view that a a broad view had to be taken of taxation issues, the one thing his contribution singularly fails to do. The reaction, however, is filling that lacuna.

cf. Seamus Coffey


‘The author advanced on RTE radio the view that a a broad view had to be taken of taxation issues, the one thing his contribution singularly fails to do.’


Taking the ‘broad view’, particularly opening up discourse on the impact of indirect taxation on lower deciles, is PRECISELY what the author has succeeded in doing.


Nonsense! Don’t you know that “the broad view” always and everywhere must include a discussion of “structural reform”? This is so regardless of the topic being discussed.

It would help if the authors clarified some points raised here, however Brendan Burgess in a Sindo piece that has a touch of Jean-Jacques Rousseau’s princess saying “Ils n’ont pas de pain ? Qu’ils mangent de la brioche!” unwittingly makes a point:

The level of indirect taxes anyone pays depends directly on how they spend their money. If you spend most of your money on necessities like rent, food and public transport, you will spend very little in indirect taxes.

If you spend most of your money on take-away food, chocolates and crisps, you will pay plenty of VAT.

If you spend a lot of time in the pub, you will give 32pc of what you spend to the Exchequer.

But if you spend a lot of time outside the pub smoking, you will give 80pc of your expenditure on cigarettes to the Government. Cigarettes are the real killer.

In his 1937 book, ‘The Road to Wigan Pier,’ on the grim conditions of the English working classes, George Orwell wrote:

“The basis of their diet is white bread and margarine, corned beef, sugared tea and potato – – an appalling diet. Would it not be better if they spent more money on wholesome things like oranges and wholemeal bread?…Yes it would, but the point is, no human being would ever do such a thing.…A millionaire may enjoy breakfasting off orange juice and Ryvita biscuits; an unemployed man does not…When you are underfed, harassed, bored, and miserable, you don’t want to eat dull wholesome food. You want to eat something a little bit tasty.”

Burgess also refers to the top effective rate but it’s interesting that in the UK, when the incremental loss of the personal tax allowance and a total loss kicks in at £120,000+ earnings, coupled with benefit impacts, the effective UK marginal tax rate is 60% (including national insurance).

@ MH

We are not in the ’30s. That there is social deprivation is uncontested. However, only evidence-based decision-making will bring the necessary improvements. The value of the contribution under discussion is that it is helping to provide the basis for such decision-making; if maybe not in the way the author intended!

re: Brendan Burgess article:

To paraphrase Mr Burgess, along the lines of Margaret Thatcher:
There is no such thing as a household, there are just people. Especially, people burdened by taxes.

America was once the shining light on the hill not just for its own citizens but for the many Irish who desperately needed a beacon of hope. It is ironic that their decline began with Ronald Regan, an Irish-American. The linked article should be read by all who have become practitioners of the fine art of begrudgery and low/no taxes in what is nominally a Christian country.
Why we should close the “unemployment industry”.
The author is Bill Mitchell an Australian economist. The content leans heavily on the US New Deal.

If I may, I would like to extend the discussion to the other positive benefits of adjusting VAT over the much discussed adjustments to income taxes.

1) Reducing VAT rates has positive cashflow outcomes for many businesses, particularly service businesses.

2) It reduces costs for a wide range of other social activities such as schools, who pay VAT, but not Income Taxes.

3) It would be of assistance for those businesses located in Border areas close to UKNI, enabling them to compete with competitor businesses in that foreign jurisdiction, reducing the amount of trade leaving the State, perhaps persuading some of the retail trade to move South from UKNI. (The location of the large shopping centres in Newry are of course there because of Irish customers, rather than the l local UK customers.)

A carefully planned reduction in VAT can have a significant influence on economic activity and should be consider in that light, not purely in the manner of an across the board reduction. Micheál’s paper is worthwhile in discussing the burden of taxation.

Michael H
I lost all time for Brendan B when
a) this and
b) his fatwah on discussion of house prices.
While he may have boned up on How to Analyse, those are pretty large balls and chains on his record.
And this comment of his
“Anyone who wishes to, can cut down their indirect taxes without government intervention – cut down on the booze and cigarettes, stop eating take-aways and take the bus instead of driving” reeks of a contempt beyond words for the lower earners.

@ Sam M

He did not complain when the reduced rate of VAT was selectively reduced to prop up Denis O’Brien’s purchase of Independent Newspapers and the restaurant trade.

@ Mickey Hickey. It would indeed be ‘nice’ if things might be as straightforward as Bill Mitchell seems to suggest. The reality is a tad more complicated – especially in the political arena.

I penned a reading suggestion to eamonn moran – its above, at 3.46 pm Aug 29th; Prasad provides the overarching insights and the explanations of the Neo-liberal agenda and its political practices. Quite unsettling they are. Ireland is just a nasty, copy-cat, mini-version of the US (and in some respects, the UK). Funny that.

The US economic decay set in after the first OPEC oil shock. [Strange how a relatively short-term supply constraint of a black, smelly liquid could derail the world’s most advanced economy.] The rot started with Ford and Carter. Regan gave it a good boost, but that ran out of steam in mid-1980s. Things free-wheeled from there on and daddy George just let things simmer. But Clinton re-boosted, and baby George applied some ‘delicate’ finishing touches.

@ Mickey Hickey: I am of two minds – whether to ‘thank’ you or not, for that link. However I loved the “barge economics” bit. Make a great cartoon. The rest is quite depressing. And as to the leading question – “Are these policies sustainable?”, the answer must be in the negative. Let’s see what happens.

That article by Brendan Burgess in the Indo was little more than a vitriolic attack on lower working classes. Very popular with many of his perspective customers though.

I would like to point Brendan to Table A3 in the report.

There is actually no evidence (this is CSO data remember) that the lowest deciles is pigging out on Takeaways, Alcohol and cigarettes. Though it is a very common stereotype.
Surely Brendan isnt basing this free advice on the basis of class based stereotypes?

The % of the lowest deciles expenditure on Food Alcohol and cigarettes is in line with all the bottom 5 deciles.
Maybe it comes as a shock but in absolute terms the people in the bottom decile are spending about a 1/5 of what those in the top decile spend on Alcohol and Fags according to the report.
His opening gambit is an obvious falsehood.
“THE Nevin Economic Research Institute (NERI) published some astonishing research during the week which has been reported uncritically in the media”
Almost every single broadsheet article I have seen has criticized the research.
In fact there was another article in the Sunday indo on the topic from a tax advisor (cant find it online) that was even more critical than Brendans article. No doubt the Indo are playing to their audience.

I also think that he and many others are misdiagnosing who the lowest income decile are. Even the author of the report thinks some could be pensioners with savings. Again pensioners on just the state pension puts them above the average lowest decile average.

“The bottom 10pc of households report having an average income of €9,800 each. Who are these people? A family of two unemployed parents and one child has an income of €19,500 between social welfare and child benefit.”

Single occupancy bedsits and apartments are considered households in this survey.

If you are a student you may have an income of 5-7 thousand but are probably getting support from your family.
Or what if you are a foreign student here to learn English living in a bedsit or small apartment. You can only work a maximum of 20 hours a week and you are most likely getting a lot of help from back home. Then there are the large number of unemployed people who are living alone those above 25 get less than the lowest income average and those below get a lot less (as low as 5,200).
Even if this group made up less than 5% of the population they could make up over 10% of the households as per the CSO definition of households.

The typical lowest decile household is not a mother and father both unemployed in social housing with kids. They are 2nd/3rd decile.

Its more likely to be a young single person either in education or unemployed living alone.

@ Sam Maguire

Brendan Burgess recommends investment in Irish bank shares one day after Lehman Brothers filed for bankruptcy and at a time when most of the big developers were not even paying interest on loans!

In Ireland with its timid media, this type of record doesn’t seem to matter.


Sure we’re not in the 1930s but we are back with long-term unemployment where the loss of a job by a person of middle age can mean permanent unemployment.

It’s a grim scenario for people who want to work – of course they can join the 227,000 1-person operations – but remote for people who have never experienced it up close.

As for the claim that “only evidence-based decision-making will bring the necessary improvements,” maybe or not?

Michael Hennigin
His record is why I didn’t much rate him but I try take each pronouncement on its merits. His ‘tude in the article however reinforced a view of someone well off the main drag of economic reality. This

@ Seamus and everyone – indirect tax costs may also be borne by works and not just reduced profitability so leading it to be more regressive

One of the most revealing things about a person is their attitude to those worse off than them. A whole pile of epic fail on display here.

@Eamon Moran.

Nice work, even if you are too kind. You have some terrific responses that expose the shallowness, callousness and plain old foolishness of many of the attacks on the Collins paper. (I have immediately spotted a fatal flaw in the study! Sure the poor are well off by medieval standards. The data’s not good enough to work from.)

Collins’ follow up finishes off the job.


You again offer a master class in the tics of rightist rhetoric.

That VAT takes a higher share of low incomes than higher ones is not news.

Here you are attempting to downplay the significance of something by saying that everybody knows it, it usually follows the author having denied that the non news was true at all and is followed by the claim that the former non news is either all for the best or inevitable given the current political and economic “realities”.

Two previous examples:

* Austerity is expansionary; sure everyone knows austerity is contractionary; austerity is simply the law.

* We are hitting the 2% inflation target; yes we are missing the inflation target in the short term but in the medium term it will be fine; deflation is not so bad is it?

The common factor is that the analysis is a way of justifying the policy choice and not a way to choose between policies.

Ireland sticks out like the proverbial sore thumb and there is no real mystery as to why. Decisions on taxation and social transfers are largely taken on the basis of political desiderata.

It is profoundly worrying that the modern right’s undisguised contempt for the wishes of the hoi polloi is excused as hard minded realism rather than condemned as a visceral anti-democratic impulse.

We are not in the ’30s. That there is social deprivation is uncontested. However, only evidence-based decision-making will bring the necessary improvements

Given your neoliberal credentials, support for austerity and ECB policy I suspect that you might be confusing the word “evidence” for the word “confidence” here.

‘Is the regressivity of the indirect tax system is driven by the bottom decile?

No. As table 7 (p18) of the paper shows, even ignoring the bottom decile the system is regressive.’ [ M. L. Collins update]

Comments are closed.