Fun and Fury, Part II

There’s been an interesting discussion going on around Ozplogistan about increasing public support for high levels of taxation and public expenditure, and the reasons for this increase. But reading a new CIS report by Sinclair Davidson, I’ve discovered, to my surprise, that it’s all my fault!

I’m not entirely alone. Peter Saunders’ introduction blames the ‘campaign of vilification and obfuscation waged by socialist politicians like Wayne Swan and socialist academics and journalists like John Quiggin ’ and the ACTU gets a mention as well. But I certainly get mentioned as prominently as anyone.

In fact, my little book[1] Taxing Times is the only academic work on tax policy that is mentioned in the piece more recent than Henry Simons 1938 book, Personal Income Taxation, which is attacked in a footnote. In fact, it appears as if Davidson is entirely unaware that the vast literature on tax incidence even exists. When he comes to estimate the impact of the GST, for example, he refers to a remark by Adam Smith that the burden of consumption taxes is roughly proportional to income and makes this the basis of his analysis.

The title of the report is Who Pays the Lion’s Share of Personal Income Tax? and, to be fair ,Davidson manages to get the answer to this question roughly right. Since upper-income earners, by definition, get most of the income, and since the tax scale is progressive, most personal income tax is paid by those with high (reported taxable) incomes. I wasn’t surprised by this, but perhaps some people might have been.

But as everyone with even an elementary knowledge of the subject is aware, the income tax is the only significantly progressive element in the tax system. Davidson quotes me (correctly for once) as asserting that, when other regressive taxes are taken into account, the tax system is roughly proportional. Davidson then attempts to take account of the GST but, in the absence of any knowledge of the literature or the relevant data sources such as the Household Expenditure Survey, makes a total hash of it, relying as I’ve noted on a passing remark by Adam Smith as the basis for his analysis. He doesn’t even mention other regressive taxes and appears to be unaware of their existence[2].

Davidson is also apparently unaware that the whole issue has been analysed properly, by Ann Harding and Neil Warren of NATSEM, who presented a comprehensive analysis of Who Bears the Tax Burden in Australia (PDF file) back in 1999.
Their key finding is:

the impact of the whole system is progressive, with tax burdens rising from 36 per cent of the total income of the
poorest one-fifth of households to 51 per cent of the total income of the most affluent one-fifth of households. The progressive effect of income tax is thus substantially, but by no means fully, offset by indirect taxes.


I’d say that this is “roughly proportional”, though judgements might differ on this. Harding and Warren show that the tax scale is roughly proportional over the middle three quintiles, but mildly progressive at the top and bottom. Certainly their findings are radically different from the picture painted by Davidson.

My claim that the incidence of taxes is roughly proportional is strengthened when tax avoidance is taken into account. Since income tax is more easily avoided than others, and avoidance opportunities are mostly available to the top quintile[3], it’s safe to assume that their actual effective tax rate is lower than that implied by the tax scales, which are used as the basis of the Harding-Warren analysis.

This, by the way, is a topic on which Davidson really excels himself. Near his conclusion he says:

Swan gives the impression that individuals with multi-million dollar base salaries are common. In the 1998/9 financial year the ATO reported only 431 individuals who earned more than $1 million. By its vagueness in defining who in particular the ‘rich’ are38 and using extremely high income individuals as examples, the revenue lobby sows confusion. If, according to John Quiggin, tax is voluntary for anyone who can afford a good accountant, then Table 2 indicates that the top 25% of income earners in Australia ‘voluntarily’ pay 64.1% of the net tax.

The quotation given by Sinclair is from Taxing Times and refers to a period in the 1970s when the Barwick High Court upheld a wide range of tax avoidance schemes and goes on to refer to the various measures that were introduced subsequently to address the problem, but Davidson makes it appear that I am talking about the contemporary situation when tax avoidance, while still significant, is much less blatant.

More significantly, Davidson has forgotten that his entire analysis is based on taxable incomes declared to the ATO. To the extent that tax avoidance is taking place, his analysis misses it completely. There could be 1000 people with an income of more than $1 million, or 10 000 for all he knows. Of course, tax avoidance isn’t easy to define or measure, and I don’t claim to have a good measure, but it’s easy to find plenty of very rich people with low taxable incomes, starting, notoriously with Kerry Packer.

Coming back to the real analysis of the topic by Harding and Warren, their data does indicate that the bottom quintile pays a somewhat smaller proportion of their income in tax than the rest of us. Of course, this is as it should be. But notice that a rising average tax rate implies a high marginal rate, and this is compounded by withdrawal of welfare benefits. If we’re concerned about incentives, this is where we need to look.

fn1. As I recall, this book sold about 500 copies. But I get little cheques for Public Lending Rights that seem to indicate that people are borrowing it from libraries. So maybe the message is getting through. Or perhaps I’ve managed to convert large numbers of Fin readers to socialism.

fn2. I’ve been noticing a decline in quality control at the CIS for some time. Although they still turn out some good-quality research, there’s been some real rubbish lately, such as the study on speeding by Buckingham that turned out to have been cribbed from the talking points of a UK lobby group. Still, this piece marks a new low – even the IPA has rarely published work as shoddy as this.

fn3. Opportunities for evasion are more evenly spread across the income distribution.

16 thoughts on “Fun and Fury, Part II

  1. The extent of tax evasion is not measurable, as it is part of the black economy.
    The extent of tax-avoidance is measurable by implication, given the differences between declared gross incomes and net taxable incomes. If this gap is so large as to flatten the net progressivity of the income tax scales then it indicates that the spirit of the law is being defied by schemes.
    Regarding the incidence of taxation and its effects on the supply and distribution of labour throughout the global economy, this matter is covered well in Does Atlas Shrug?.

    Since the introduction of the income tax in 1913, controversy has raged about how heavily to tax the rich. Opponents of high tax rates claim that heavy assessments have negative incentives on the productivity of some of our most talented citizens; supporters stress the importance of the rich shouldering their
    “fair share,” and decry the loopholes that permit many to escape their obligations. Notably absent from this debate is hard evidence about the actual impact of taxes on the behavior of the affluent.
    This book presents evidence by leading economists of the effects of taxes on the formation of businesses, the supply of labor, the form of executive compensation, the accumulation of wealth, the allocation of portfolios, and the realization of capital gains. Among its findings are that the labor supply of the rich remained unchanged in the face of large tax cuts in 1986, and that in late 1992 executives exercised billions of dollars’ worth of stock options in order to beat the tax increases expected in 1993. The book also presents a history of efforts to tax the rich, a demographic snapshot of the financially affluent, and a road map to widely used tax-avoidance strategies.

    Sinclair Davidson’s report contains no reference to this report, which is a oversight.
    His definition of middle income as extending from 30,000 to 80,000 pa is a little expansive, to say the least.
    The ABS reports that, as of Feb 2004, Average Weekly Earnings, for part and full time employees, are at $751.30.
    Assuming that the typical household earns 1.5 X AWE that would give an average weekly income for each household of ~$1100.00. Probably closer to $1,000 given the unequal distribution of income.
    So middle income households would probably stretch from $30,000 to $70,000 per annum.

  2. Actually, the original (and justified, in the absence of a social security safety net) objections to income tax were largely centred around its impact on precarious incomes – recall that this was in the early 19th century, when that was a more technical term, and was being used accurately.

    As for not being able to measure the size of evasion because it is part of the black economy, well, no, not directly, but there are statistical estimator techniques with known reliablity. These are used to assess how many proofing errors have not been detected, how many stars there are in the galaxy, or even how many crimes have not been detected let alone solved. I once heard that the British police researched how many murders they had overlooked and hastily left it unpublicised when they found out how large it was.

  3. hi john,
    like the analysis. found this a bit funny though:

    “I’d say that this is “roughly proportional”, though judgements might differ on this. Certainly it’s radically different from the picture painted by Davidson”

    i agree with your tone, but 36% is not “roughly” the same as 51%. as a matter of rhetoric (where are you deidre mccloskey) i think that you are better off granting the progressive nature of the australian tax code and celebrating that as a sign of a socially conscious/oriented nation. otherwise, it looks like you are apologizing for/or sensitive about this particular australian trait.

  4. few points of disagreement

    1) the GST is not regressive. A broadly based consumption tax is the only proportional tax (the effect of having fresh food excluded notwithstanding).

    2) I think you’re being cheeky to call the Harding findings “broadly proportional”. I think a more objective assessment of the findings would say “marginally progressive”

    3) you say that a tax system should “of course” be progressive. There is no “of course” about it, but I accept that it’s your view

    4) To jack – there have been many many studies on the incentive effect of income taxes. I think it fair to say that standard economic opinion is that there is a negative supply effect, though the size of this effect is still an issue of quite some debate.

    Having said all of that – I should add that Q and I do agree that most taxes are regressive. This causes me no moral pain because I am a strong advocate of reducing all taxes… not just switching taxes around. Indeed, I am one of the few who believe in abolishing the most regressive taxes – the sin taxes (smoking, gambling, alcohol etc). Further, we agree that the incentive effect on the lowest tax payers is a serious problem. I would say the most serious economic problem that Australia faces today. I imagine our proposed solutions would differ though… 🙂

  5. cas and john: fair enough, I’ve changed the summary of H&W to something more defensible.

  6. PML, I’m quite familiar with the literature on tax evasion, measurement of the hidden economy and so on. There are techniques, but they are unreliable and fraught with difficulty.

    I see no reason to revise my statement that avoidance is hard to define and measure – the same is true for evasion.

  7. As an Australian exile in Mexico, I can tell you that the numbers for Mexico in Table 1 are totally off base. Anybody with half a brain can see that by looking at the numbers for Mexico. You don’t need to be an economist. They make no sense.

    While I lived in Australia, when I got a promotion, I could package my salary. As my salary went up, my ABSOLUTE amount of tax paid went down! And all of this was perfectly legal.

    I do not see rich folks from Australia are in a stampede to leave Australia and go to live in some of the other low tax countries mentioned. This tells me that the rich are not exactly hurting that much!


  8. John,
    You should be extremely cautious about relying on income figures collected in the Household Expenditure Survey. The ABS is not exactly thrilled with collecting household income in this particular survey, but does so to satiate the unreasonable demands of social planners.It is essentially an expenditure survey, largely for weighting the CPI, although it does try to satisfy the marketing fraternity as well.

    One of the glaring anomolies of the HES in its income data, is that a large number of low income households are spending more than they earn. This is clearly rubbish in the medium term, but reflects the inadequacy of a snapshot approach to income. True incomes for these households have to be followed up over time. As a result, income and expenditure conclusions drawn from the HES should be highly qualified. The ABS would probably say, highly suspect.

  9. observa, I’m well aware of all this. The usefulness or otherwise of the HES has been the subject of vigorous debate, in which I’ve participated, and various alternatives have been tried. This is the kind of thing that real tax economists do for a living.

    My point was not that Davidson should have used the HES but that he doesn’t even mention it, or any other relevant data source.

    To take an analogy, there are all sorts of problems with GDP as a measure of living standards, but somebody who attempted to do an international comparison of living standards by ignoring GDP data and relying on a couple of quotes from David Ricardo would rightly be dismissed as an amateur.

  10. Talking of Ricardo, our own Dave Ricardo should really love Davidson’s footnote 37

    “37 The last year for which this particular data are available.”

  11. I wasn’t suggesting that the statistical estimators were good enough for what we need but that we had them rather than nothing at all and that we knew how much weight we could put on them.

    As for unreliable measures, I liked the story of what happened when VAT (GST) came to Sicily. The total land area being claimed for rebates somewhat exceeded the total area of Sicily, even allowing for cliffsides etc.


    Although this is slightly tangential to the topic, I heard Sinclair Davidson on Life Matters (Radio National) this morning debating Anne Harding on the tax issue. Verdict: Davidson may be a smart cookie, but he’s not a communicator!

    However, the point I’d make is that Life Matters saw fit to put Davidson up against an opponent; whereas the day before when interviewing an Australia Institute researcher about a paper she’d just released on government intimidation of charities, the other person Life Matters interviewed was, surprise surprise, supportive of the Australia Institute position.

    This is only a sample of one, but typical I think of Radio National’s approach to balance in public debate.

  13. Davidson may be a smart cookie, but since it doesn’t show either in print or, apparently, in person, it’s hard to see how we can know this. I’ll have to listen to the show.

    On Life Matters, I’ve usually had an opposing view when I’ve been on – Peter Saunders himself on at least one occasion I think, though it may have been a different RN program.

    I think that Harding & Warren are as close to the “neutral expert” category as you get in this field.

  14. My use of the term “may” was deliberate – the only evidence of Sinclair’s brilliance to date is the letters after his name and, as I’ve noted previously, they are neither a necessary nor sufficient condition for sensible argument.

    On the bias issue, my point is that, in my experience (or, at least, perception), one rarely hears right-wingers being given an uncontested chance to lay out their views on RN, whereas this opportunity is commonly provided for lefties. That Q’s appearances on RN have generally been with opponents does not alter this point – there are of course many interviews on RN that set right against left.

  15. In the debate over the incidence of taxation on particular levels of income, it is largely espoused that high income groups are responsible for the majority of tax minimisation/evasion. IMO the flip side of this coin, is it should also be recognised that low income earners are the largest participant group in the black economy. I think(aside from the income snapshot collection problem)the HES expenditure results for low income households reflects this.

    I personally know of a SPB beneficiary, who owned her own home and car and privately schooled a son on supplementary income from private cleaning. With an adult son, she now works full time. Another male invalid pensioner I know of, lives with his aged parents in a lined garage and would earn probably another $40k per annum out of this growing MJ hydroponically. SA is a major backyard supplier of MJ interstate. These sorts of incomes will never be reported in HES, although the two week expenditure diaries of these participants,as well as the recall of major durable purchases, would show the true nature of their incomes and lifestyles.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s