The direct and indirect tax debate
By Sean Whelan, Economics Correspondent
Since we posted this blog Micheal Collins of NERI has posted a response to some of the issued raised here and elsewhere (see the IrishEconomy.ie thread especially). It’s worth including all of it here, especially as it is in the form of a Q&A, so it is user friendly. As we said, it has certainly kicked off a good debate. The new NERI Q& A is further down the page, after the original blog post.
The Nevin Economic Research Institute seems to have kicked over a hornets nest with its latest version of a working paper claiming that top and bottom income deciles pay about the same proportion of their income in tax.
The premise is that while income tax is very progressive in Ireland, indirect taxes such as VAT tend are regressive.
Using data from the CSO’s Household Budget Survey economist Micheal Collins says the combined effect of direct and indirect taxes results in three income deciles – 1, 9 and 10 – paying above average levels of tax. Because decile 1 is the lowest income group, NERIs research suggests the total tax system is regressive.
The implication is that reducing or otherwise modifying indirect taxes such as VAT and/or excise would disproportionately help raise incomes or purchasing power for households at the bottom end of the distribution.
That may well be correct – most economic opinion here and abroad views indirect tax as broadly regressive.
But is it regressive to the extent NERI claims? Or is the amount of tax attributed to this decile an outlier that gives the NERI chart its distinctive “U” shape (actually more of a Nike Swoosh-shape)? And if its an outlier, does it stand up?
One correspondent wondered if net contribution to exchequer would be a more meaningful figure to take account of social welfare payments.
And then there is the suggestion that the average household direct tax charge is 13.6% and the average household indirect tax burden is 10.36%, giving a total tax charge of 23.95%. It certainly doesn’t feel that way.
Earlier versions of the findings attracted criticism (not all of it temperate) about the assumptions used to arrive at the direct tax figure in particular. Brendan Burgess of AskAboutMoney posted the following:
The Nevin Economic Research Institute has launched their Quarterly Economic Observer
Their key point is:
“Currently the top 10% of households (above €109,000 gross income per year) have an average effective tax rate of between 22.5% and 27.5%”
Anyone lucky enough to be earning €109,000 per year must be shocked to see that they are paying only 22.5% in tax.
So how does the Nevin Institute come up with this figure? It takes a lot of statistical wizardry to show that the top earners in this country pay an effective tax rate of 22.5%.
Trick no.1 Talk about households, not individuals.
So a household is all of the people living together in one household. So this includes:
- a single person living on their own, earning €109,000
- A couple where each earns €55,000
- three nurses sharing a house who earn €37,000 each
Trick no.2 Redefine gross income
Add in the following:
- Employers’ PRSI
- Child benefit
As a result, a couple have to earn only €45,000 each to be assigned a gross income of €109,000 which catapults them into the top 10% of households.
Trick no 3 – Double count pension contributions as income
- Add the employers’ pension contribution to gross income
- Don’t allow the employees’ contributions as a deduction
- But when the person draws down the pension in retirement, show it as gross income.
Pension contributions whether contributed by the employer or the employee are deferred income and should only be taken into account once – when the person receives the pension income.
A comparison of the true effective tax rate with NERI’s calculation
|Neri’s gross income
couple earning €45k each
|Gross salary per payslip||€90,000||€90,000|
|Less employees’ pension contributions||€ 4,500|
|Add employers’ pension contributions||€ 9,000|
|Add employers’ prsi||€ 9,675|
|Add child benefit for two kids||€ 3,120|
|Taxable income/ NERI’s gross income||€85,500||€111,975|
|Effective tax rate||42%||22.3%|
Trick no. 4 Ignore the Revenue’s own data on effective tax rates
The Revenue Statistical Report for 2011, gives the data for income tax for 2010. (Extract attached to this post)
For single people earning between €100k and €150k
|Income tax paid||€239m||€150||89m|
As tax credits, have been reduced, it’s likely that the effective tax rate has risen marginally.
So the effective total tax rate for 2013 for a single person earning between €100k and €150k is
Trick no. 5 Don’t specify how you would raise the “effective tax rate from 22.5% to 24%”
For example, don’t suggest any of the following which are implied by the way the figures are presented:
- The married tax credit should be abolished. Married couples should be treated as if they were a single person, in the same way as we present our data.
- All people living in a house together should be taxed as one person i.e.they should have only one set of tax credits between them.
- The tax relief on employees’ pension contributions should be abolished but pensions in retirement should continue to be taxed.
- The tax relief on employers’ pension contributions should be abolished also.
- Employees should be taxed on the PRSI paid by their employers, and any resulting benefits such as Jobseekers Benefit or Maternity Benefit should continue to be taxed as well.
- Child Benefit should be taxed.
Seamus Coffey of University College Cork economics department posted the following response to the original working paper, finding the total tax system is progressive. He also has a chart including state transfers as a proportion of income:
- NERI looked at the combined effective rate of direct and indirect taxes on Irish households along the income distribution. The headline results are summarised in this chart.
- From the fourth decile up the results indicate that the Irish tax system is progressive. The first three deciles run counter than this though with the first decile notable in particular because it has the second highest reported effective tax rate.
- The lowest contributions are in the third and fourth deciles and we have previously looked at the composition of the households in those deciles (albeit using the SILC rather than HBS as used by the NERI study).
- So what can we say about these estimated tax burdens. The following table gives the estimated nominal amounts paid by each decile.
- The estimated average tax burden is €12,900 per household. According to the Census taken in April 2011 there were 1.7 million households in the country. Thus the total amount of tax included in the effective tax rates is €12,900 x 1.7 million = €21.9 billion (of which €12.5 billion is “direct” and €9.4 billion is “indirect”). That is a lot of tax but it is around half of the total tax actually collected in 2010. The 2010 tax take is summarised in this table.
- Of course, the incidence of taxation is incredibly difficult to determine. Direct taxes are deducted from income but even that is not enough to isolate the economic incidence of the tax. Do employers need to pay a higher wage to attract workers from a abroad in locations with high income taxes. This is an issue in various soccer leagues. Determining the exact incidence of indirect taxes is almost impossible.
- If a household buys something for €123 that includes VAT at 23% then the amount of VAT in the price is obviously €23. However, this doesn’t mean the purchaser paid the full amount of VAT. It depends on what the price would be in the absence of the VAT.
- If the VAT was eliminated the price might fall to, say, €110. It is incorrect to say the household has paid €23 VAT when the price absent a VAT would be €110. In this case the retailer, or some other element along the production chain, has absorbed part of the VAT.
- The NERI report covers €5.7 billion of VAT (€3,360 x 1.7 million) as opposed to the €9.9 billion actually paid but it is not stated why this gap is present. It could, in part, be because of VAT paid by businesses not fully passed on in the price. Of course, this will be reflected in lower profits and dividends from the business so the burden of the taxation will ultimately be borne by households as is the case for all taxes. Some of those households may be outside the country as with Corporation Tax with around 75% of the total paid by foreign-owned companies.
- There will be many small businesses who would regard the rates, excise duties and other taxes which they might not be able to pass on through prices as coming out of their income. Of course, there will also be many cases where rates, duties and other taxes are put into the price but are not explicitly listed. However, in the sense that these may affect all households equally it may not alter the overall results. Then again they may not.
- Returning to the household level it is worth looking at the income and expenditure figures for each decile. The tax rates shown in the first chart above are as a percentage of gross income while the dominance of indirect taxes for the lower deciles means that the amount of tax paid is, in the main, a function of expenditure. This table from the report summarises these.
- One of the notable features is that average expenditure of households in the lowest decile is almost two times greater than their disposable income. This is also true for the second, third and fourth deciles though not to the same extent.This is not an unusual feature of household surveys and the following is noted by the CSO in their publication of the 2009/10 Household Budget Survey:
- There are many reasons why expenditure may exceed income in lower income decile households and this is a common experience internationally in income and expenditure surveys. Households with recently unemployed household members may draw on savings to maintain their expenditures. Self-employed consumers may experience business losses that result in low incomes, but are able to maintain expenditure by borrowing or relying on savings. Third level students may get by on loans or savings from summer employment, retirees may rely on savings and investments. In addition, across all deciles there may be an under-reporting of certain categories of income (e.g. shadow economy employment income).
- On the income measure in the HBS the CSO note that:
- The HBS [.] calculates income on the basis of the “current income level” of the individual without adjustment for employment activity over the year in question.
- The HSB is a weekly survey rather than an annual one with respondents reporting their income and expenditure over a two-week period. Is it appropriate to calculate an effective tax rate using expenditure now when the income may have been earned previously.
- In the 2009 SILC the average disposable income of households in the lowest income decile (as measured over a year) was €11,000, around 12% higher than the equivalent figure in the HBS. For the top decile the SILC has a figure of €118,300 versus €119,500 in the HBS, a difference of 1%. The HBS may not the optimum instrument to measure income but it is good at what it is designed to measure: expenditure. Here are the expenditures, estimated indirect tax burdens and effective tax burdens as a percentage of expenditure by decile.
- Although VAT has some progressive features with most necessities zero rated it can be seen that the proportion of expenditure accounted for by indirection taxation is inversely related to expenditure. The greater the level of expenditure the lower is the proportion taken in indirect taxes. This is likely due to consumption patterns, particularly in relation to high excise duty goods, and also the disconnect between some indirect taxes and expenditure such as the television license, credit/debit card levies and annual motor tax to a certain extent.
- Although there may be difficulties measuring it there can be little doubt that indirect taxation in Ireland is regressive. Whether it is sufficiently regressive to offset the very high progressiveness in direct taxation is less clear cut. Finally, the CSO also give a breakdown of the income by direct income and state transfers for each decile which can be used to infer that Ireland’s overall tax and transfer system is progressive.
As is so often the case, the headline findings are abstracted from an abstract that masks some pretty difficult data wrestling. The data in question is that CSO Household budget survey.
The most recent version was published in 2012, and is based on fieldwork carried out in 2009/2010.
Obviously a lot of changes have been made to the tax code since then, especially in Budgets 2011 and 2012.
But the CSO only do this survey every five years (that seems set to change to an annual survey after 2017). The author notes that while it is the only comprehensive source of household expenditure data, it is “far from a perfect measure”.
The paper does try to address two of the post survey changes to indirect tax – the rise in VAT from 21% to 23% in Budget 2012 (which it finds to be regressive) and the cut in VAT to 9% for tourism and hospitality related sectors (which it finds to be progressive).
Micheal Collins says his aim with the paper is to stimulate debate on the impact of indirect taxes. He says the debate on tax is too focussed on the rates of income tax and USC.
That may be true, but trying to make indirect taxation a greater part of the political tax debate will, one suspects, take a good few years – and a lot more data.
The response from the Nevin Economic Research Institute
Where did you get this data?
The analysis is based on the CSO’s most recent Household Budget Survey (HBS). The HBS is collected to establish data on the composition of the ‘typical basket of goods’ that consumers spend – the basis of the inflation data published each month. In that regard it offers a good insight into the expenditure pattern of households.
The data and the survey are discussed over pages 7-11 of the paper.
Is there a lot of VAT missing?
No. When you compare the estimated average VAT collected in the paper to the overall level collected from households they match quite well. To quote the paper (p11):
“As a means of assessing the robustness of the modelled indirect taxation, table A13 of the appendix compares the calculated total VAT tax take (the average household level times the number of households) with the exchequer revenue from VAT. Overall the modelled VAT collected from households equals just over €5.5 billion representing between 54.5% and 56% of the exchequer VAT collected in 2009 and 2010. Estimates from the European Commission for the period 2000-2011 suggest Irish households contributed on average 49% of the total VAT tax take; with the remainder coming from investment (28%), industry (18%) and Government and non-profits (2%) (2013:60-61). The EC household VAT estimates for 2009 and 2010 were 51% and 53% respectively suggesting the modelled indirect taxation outlined above, and used in the remainder of this paper, offers a good representation of the indirect taxation experience of households”.
See more on this on pages 11 and 38 of the paper.
A forthcoming paper (the subject of a NERI research seminar on Sep 10th) also tests the modelled indirect tax estimates against Department of Finance estimates of additional revenue from indirect tax changes. Again, the modelled estimates are found to offer a good representation of the indirect taxation changes.
Why are there gaps between the levels of income and expenditure in the lower deciles?
As the paper shows (page 7 Table 2) there are households in the lower income deciles who spend more than they have in income. While there are small differences for deciles 3 and 4 – probably linked to the structure of the survey and its income sampling period – there are more notable differences for the bottom two deciles. In the Irish and international income distribution literature this is not unusual. To quote the paper (p8):
“It also reveals that expenditure exceeds disposable income for the bottom four deciles, most notably for the bottom quintile, reflecting the composition of these households (for example pensioners who may also be living on past savings, the temporarily unemployed and students), their difficulties in making ends meet and the structure of the HBS which compiles its income and expenditure data on a current basis over the two weeks of a households participation in the survey”.
A proposal made by a respondent to this paper at the recent International Institute of Public Finance conference raised the possibility of inferring income levels by modelling them using other household variables and characteristics. This might help in giving a better income picture for lower deciles, although it would be unlikely to radically change things.
Why not compare the levels of taxation with disposable income?
As the paper is looking at both direct and indirect tax contributions, a comparison with one income definition is required. Reflecting the way in which we consider income taxes (as a % of gross income) and given past examinations in the literature, the gross income measure is used. In fairness, this seems the correct baseline for making comparisons. However, for completeness, the paper does include tables with comparisons against both disposable income and expenditure – see tables 14a, 14b, 16a and 16b (pages 39-41). To quote the paper (p18 regarding data on table 7):
Tables A16a and A16b in the appendix present another set of comparisons benchmarked against equivalised disposable income and equivalised household expenditure. The disposable income results are broadly similar while those compared to expenditure mitigate some of the regressivity reported above.
How is the income and direct tax data you use calculated?
The CSO provide this data in their published HBS report and in the microdata file (from the Irish Social Science Data Archive) which was used in undertaking this research. There are more details on this on pages 11 and 12 of the paper.
The direct taxation estimates include income taxes and employee PRSI contributions. The estimates from the HBS line up well with those I established last year using the more detailed SILC data set (the principal Irish income data set). See tables 4 and A4 of that paper – the paper and the methods in the paper are all outlined here. That paper also compares the estimates with the published Revenue Commissioners data on effective tax rates and finds them to be a good fit.
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.
Does the research imply a need for indirect tax cuts in the next Budget?
No. The research is just that – research on the shape of the tax contributions people make. The paper takes no policy positions. To quote the paper (p2 and p3):
In establishing these estimates, the paper aims to provide a more comprehensive understanding of the distribution and composition of household tax contributions. Understanding the overall shape of household tax contributions offers a firmer basis for considerations of policy options, or critiques of previous policy changes. Limitations in this understanding have been obvious in recent policy considerations…
Similarly, there has been limited consideration of who gains or loses from increases to excise duties, amendments to insurance levies or extensions of various indirect tax changes. While this paper does not address each of these issues, in establishing an up-to-date baseline understanding of the total tax contribution distribution, it provides a basis and context to begin to consider these issues. Such a framework should also assist in the inevitable future considerations of where tax reductions for individuals/households might best be targeted. Perspectives on such choices are likely to be different when judged across overall household tax contributions rather than solely on income taxation.
Should you not take differences in household composition into account?
Yes, and the paper looks at both the household level and the equivalised household level (adjusted for household size and composition). The latter are the figures highlighted (although the findings are similar for both) and used as the baseline for future policy examinations.
To quote the paper (p11):
Finally, the analysis considers taxation patterns first at the household level and then at the equivalised household level. The former provides an insight into the proportion of household income/expenditure that is consumed by indirect taxes. The latter adjusts the data (both expenditure and income) to account for differences in household size and composition. The analysis uses the national equivalence scale with values of 1 for the first adult, 0.66 for each additional adult (aged 14yrs+) and 0.33 to each child aged less than 14 years. Following equivalisation, households are ranked by gross income and divided into deciles. These equivalised household deciles are used for policy simulations later in the paper.
The unadjusted household figures are in table 6 (p18) and those adjusted to take account of the composition of households to give a per adult equivalent figure are in table 8 (p19).
Are you sure the Irish income tax system is progressive?
Yes. It is clearly a progressive tax system – as it should be as that is how it is designed (see table and chart on p 12). It is progressive as when income rises people pay more and vice versa. Being progressive does not mean it could over time become more or less progressive; that is a policy choice for years to come. One that I hope should be considered in the context of the broader taxation system and not just income taxes.
Are things likely to have changed since the HBS 2009/10 data?
Yes, since then both direct and indirect taxes have increased. There have been direct tax increases via the USC and PRSI thresholds and some small decreases via USC changes. There have been indirect increases via VAT rate increases, excise increases, insurance levies and new charges for property and water; there was also a VAT decrease for tourism related items. Indeed, we have closely watched the direct increases, but reflecting the starting point of this paper, the indirect ones have received limited consideration. It is fair to say that relative to 2009/10 people across the income distribution will be paying more in taxation.
Future work will address some of these issues. Most immediately will be two NERI working papers, the first on recent VAT changes and the second on the distributive impact of a suite of possible indirect tax changes. These papers will be released during September 2014.
The link to the full paper is http://www.nerinstitute.net/research/total-tax-contributions-of-households-in-ireland/