Government Senators do not necessarily agree with the emphasis placed on some of the information contained in the body of the Select Committee's report. We therefore have provided additional evidence contained in this separate report to support our conclusions.


This Select Committee on A New Tax System has terms of reference to inquire into the Government's tax reform plans as set out in the policy document Tax Reform: not a new tax a new tax system (ANTS).

This report covers the first stage of the Senate Inquiry, which has examined the views of economic modellers, and whether their evidence supports the case for tax reform. As set out in this report, their economic evidence supports the Government's position of tax reform providing benefits for Australia.

2. The Government's tax reform policy

ANTS is a comprehensive policy approach to tax reform which brings together

reductions in personal income taxes totalling $13 billion a year

better links between the tax system and the social welfare to remove poverty traps and increase incentives to work

increased pensions and government allowances

the introduction of a broad-based goods and services tax (GST) to replace wholesale sales tax and 9 other indirect taxes; and

a more sustainable approach to Commonwealth-State financial relations.

In establishing the need for tax reform, the Government identified the problems of Australia's current tax system and set them out in detail in the Treasurer's 1998 publication `The Australian Taxation System – In Need of Reform' and in ANTS.

Without reform, the Commonwealth will become increasingly reliant on income tax directly levied on individuals and companies and the tax revenue of the States will become even more inadequate and more inefficient.

The existing taxation system is broken.

The existing taxation system is unfair

- average wage earners will be paying 43 cents for every additional dollar in 1999 and
47 cents in the dollar soon after 2000

- high marginal tax rates punish incentive and encourage minimisation

- the wholesale sales tax system is illogical, out of date and costs jobs by taxing exports and investment

The existing taxation system is complex

- Sales tax is difficult to understand and very costly to administer and comply with, especially for small business

- 7 tax rates – 0, 12, 22, 32, 37, 41, 45 – lead to absurd classifications

- causing business, the ATO and the Courts to spend large amounts of time and money deciding which goods fit into the various categories

The existing taxation system is outdated

- it was designed and introduced in 1930 and has failed to keep pace with changes in the economy

- now, compared to the 1930s, services are a larger part of economy so a goods only tax base distorts investment

- complex multiple rates and classifications were introduced as emergency wartime measure and it was stated at time that this approach would never have been contemplated in peace time

- very few countries still operate a WST

Social security beneficiaries face high effective marginal tax rates, complexity and unfairness

- the current social security system creates poverty traps and disincentives to work

- a family can be worse off by entering the workforce

The sales tax base is shrinking

- because WST is a tax on some goods and no services, it is levied on the shrinking part of the economy because services have become the increasingly dominant economic activity

- WST revenue has only been maintained through rate increases which accentuates the division between taxed goods and untaxed services.


The Government set out in ANTS why the tax reform package will deliver substantial long-term improvements in the operation of the economy, to the benefit of all Australians.

Improvements will be reflected in higher economic growth as a result of stronger, more productive, investment which, together with the lowering of industry costs, will yield better export outcomes. Such changes will be crucial in relaxing the balance of payments constraint that has for so long held back Australia's growth performance. The combination of higher growth and improved work incentives will deliver more jobs and lower unemployment.

Key elements of the package that will drive higher growth are:

The Committee heard evidence from the Treasury, the Melbourne Institute of Applied Economic and Social Research (including evidence concerning distributional effects which is covered later in this report) and from Mr Geoff Carmody of Access Economics. The Committee also commissioned its own research from Professor Peter Dixon of the Centre of Policy Studies at Monash University and Mr Chris Murphy of Econtech.

In general terms, all the economic evidence provided to the Committee supported the Government's view that the tax package will provide substantial benefits to the economy.

Witnesses provided their assessments of the macroeconomic impacts on measures such as economic welfare gains, GDP growth, CPI effects and employment based on a combination of economic theory, models and judgements.

The breadth of ANTS, and the inclusion of measures affecting indirect taxation, direct taxation, government benefits and the fiscal balance, makes the task of modelling the package extremely difficult. Before commenting on the results of the private sector modelling it is instructive to examine the Treasury approach to estimating the macroeconomic effects of ANTS in the development of the package. In the evidence of Dr Henry

We were interested in the qualitative picture primarily: that is, is it likely that this particular package will lead to a higher level of gross domestic product over time than we would otherwise have? Is it likely that this package would lead to higher employment levels over time than what we would otherwise have? Those are qualitative questions, and for those questions we were comfortable with developing a qualitative answer.

Our secondary interest was in the question of what might be the impact on revenue of these sorts of behavioural effects: that is, if we are comfortable with the idea that gross domestic product will be higher, what might be the impact on income tax revenue and on GST revenue of that higher level of gross domestic product? [1]

We reviewed the modelling work that had been done and the sorts of estimates that people were producing for increases in gross domestic product. We reviewed all of that quite closely. We then had a discussion about what might be the most plausible range of estimates, and then we fixed on a number very much at the conservative end of that plausible range of estimates. [2]

The Treasury approach to assessing the impact of the Government's tax package accords with the views of Professor Dixon who, in evidence to the Committee on several occasions, indicated that emphasis should be placed on material underlying modelling outcomes and not necessarily the outcomes themselves.

The Chair paraphrased Professor Dixon in the following terms

…he said "Pay less attention to the numbers; pay more attention to the arguments as to how we arrived at these figures.” [3]

Gross domestic product

All modellers found that the Government's tax package would increase Australia's output or GDP.

Professor Dixon was at the bottom end of the range with his long term estimate of a 0.2% increase. The Committee was informed, however by Econtech, who reviewed the MONASH outcomes, that assumptions in MONASH about the composition of GDP and capital stock "… makes the Monash estimate of a GDP gain of 0.2 per cent misleadingly low when assessing the gain in living standards from ANTS using GDP…". [4]

Econtech's estimate of a 1.8% increase in GDP sits closely with the Melbourne Institute of Applied Economic and Social Research who expect GDP to increase by 1.7% as a result of the tax package.

Access Economics forecast a higher benefit to GDP, stating their view that the tax package will lead to a 2.5% increase in national output.

While the Treasury did not provide a point estimate to the Committee, they indicated in evidence that cumulative GDP growth of 0.5% by 2002-03 was implicit in the tables contained in ANTS which set out the fiscal impacts of the package. This is consistent with the long term outcomes mentioned above.

Export performance

All modellers estimated a beneficial impact on exports.

Professor Dixon again provides the lower bound for estimates of export growth, predicting an increase of 1.0%. The Melbourne Institute of Applied Economic and Social Research then comes in at a 2% increase. Econtech estimates a 6% increase.

The benefits to exports arise from the removal of taxes on business inputs and because exports are GST-free. That is, the tax package lowers the cost of export production. Given that there is a small appreciation of the exchange rate, these positive export outcomes demonstrate the benefits to exporters of reducing the cost structure of Australian industry.

The reason for the difference between Monash and Econtech is that Monash have lower export demand elasticities and a smaller increase in export supply arising from Monash's smaller capital stock and hence production potential.

Consumer price index

In the case of the CPI, Professor Dixon assumed a 2.19% CPI impact for the purposes of modelling, ie he used an assumption as an input to his modelling rather than the model producing an estimate for the CPI impact. As a result, the MONASH model provides no assessment of the inflationary impact.

The increase in the cost of living associated with the introduction of the tax reform package is estimated by the Treasury to be 1.9 per cent in 2001-02.

As set out in ANTS, however,

It needs to be noted,… that the 1.9 per cent CPI increase excludes the impact of the tax package on tobacco prices. This reflects the Government's view that the impact of the GST on tobacco prices should not, for public health reasons, be offset by income tax cuts and/or increases in social security payments. The 1.9 per cent CPI takes account of the new First Home Owners' Scheme's offsetting the impact of the GST on new house prices. [5]

In evidence given to the Committee, Treasury confirmed that ANTS had referred to a (first year) 2000-01 CPI impact of around 2.5%. Further, the Treasury indicated to the Committee that the CPI impact in 2002-03 was estimated at 1.8%.

The Melbourne Institute indicated in evidence that their estimate of the CPI impact was “…2.44% after three or four years, and then it will move down further…In the longer run the price will come down towards the 1.9 per cent.” [6]

The Econtech estimate of long term CPI impact is 0.9%. This lower estimate arises because Econtech incorporates second round saving effects from financing costs and businesses changing their input mix to take advantage of larger savings.

The debate about whether the CPI impact is an appropriate methodology for measuring the price impacts of the Government's tax package is raised later in the report.

Economic welfare

Economic welfare is a measure to estimate living standards. In modelling the Government's tax package Professor Dixon initially estimated a figure of negative $30 million. In his paper of 25 January Professor Dixon described this as “…the long run resource allocation gains flowing from the proposed tax changes will be negligible…” [7]

In the review of Professor Dixon's work, Econtech found that the formula used by Monash to calculate the welfare gain was an approximation only and would likely lead to an under-estimate. Because of this approximation error, Mr Murphy stated in his review of the Monash work that

… all results for the welfare gain contained in the Monash Report are under-estimates…From last-minute advice from Monash the understatement error appears to be about $100 million to $150 million on a Monash Model basis. [8]

It was established in evidence that Econtech's MM303 model is clearly superior to the Monash model in measuring welfare gains. Its estimate of the welfare gain as a result of the tax package is $607 million. Mr Murphy indicated that this result is conservative because "…MM303 allows for only four out of six sources of efficiency gain from ANTS…" and "…it is…difficult to model the efficiency gains from abolishing business stamp duties and bank account taxes…" [9].


Monash and Econtech use long term computable general equilibrium models which assume that employment is held constant in the long run. Therefore they do not forecast long term employment outcomes.

In the case of the Monash model, it is set up to follow the transitional path of variables including employment.

In the central case modeled by Professor Dixon, "…employment will be stimulated in the short-run by about 30 000 jobs…" [10]

In research undertaken by the Melbourne Institute, they indicated that modelling results on the basis of no change in pre-tax real wages, the tax changes could increase the number of jobs by 50 000 [11].

Access Economics, in their AEM Model Forecast Report of December 1998 indicate that as a result of the tax package there is an additional 190 000 jobs created in the long term, consistent with their view of increased GDP as a result of the tax package.

The Committee also requested some work on the employment effects if it was assumed that workers bargained for wages ignoring the greater disposable income they have as a result of the tax cuts.

In the Government's view, there is no case for such increases as wage earners will be more than fully compensated for consumption price increases by income tax cuts and increases in government benefits their real take home pay will actually rise.

In a continuing low inflation environment and given the enhanced credibility of monetary policy flowing from the Reserve Bank's independence, the Government considers it will be able to continue a policy framework that will accommodate some or all (depending on the circumstances at the time) of the small one-off increase in consumer prices but not to permit further flow through.

Against this background, and the improved cost structures inherent in the package, the introduction of a GST is unlikely to lead to an increase in inflationary expectations and on-going inflation.

This assessment is supported by international experience.

Econtech in its report of 14 February 1999 said

"The chances of ANTS producing a wages blowout are remote. This assessment takes into account the generally favourable experience of other countries in introducing this type of tax reform, and that Australia has the advantage of having available a Budget surplus that is used to boost income tax cuts to over-compensate workers for the CPI effect of ANTS." [12]

Fiscal effects

The tax reform package has been designed to be consistent with the Government's medium-term fiscal strategy.

The package will be implemented progressively from 1999-00 and will reduce the size of the Commonwealth surplus projected for that year and subsequent years. The package has a significant fiscal cost (somewhat less than one per cent of GDP annually) but will bring substantial benefits to the operation of the economy and to the sustainability of both Commonwealth and State government finances. Reflecting the considerable improvements in Government fiscal policy in recent years, the tax package can be accommodated while retaining sizeable Budget surpluses as required by the fiscal strategy.

The budgetary position of the States will, over time, be enhanced considerably by the package. After a transitional period in which the Commonwealth has committed that the States will be no worse off in budgetary terms, the States will be in a much stronger position, reflecting that GST revenue will grow at an appreciably stronger rate than the Commonwealth grants and State taxes that it is replacing. The Treasury predictions show that State Governments should benefit by $370 million in 2003/4 and by $1.25 billion in 2004/5.

A fundamental objective of this package is to halt the erosion of indirect tax revenue base.

The tax mix in the Australian economy has been moving over the past decade towards a higher share of direct taxation and a lower share of indirect taxation. For example, in 1987-88, indirect tax represented 6.9per cent of Gross Domestic Product (GDP) compared with 5.6 per cent in 1997-98. The direct taxation of individuals and companies increased from 16.9 per cent of GDP to 17.5 per cent over the same period.

As revenue from indirect taxes has declined, more of the burden has had to be carried by taxes on income; principally by wage and salary earners. The share of indirect tax as a proportion of total revenue to the Commonwealth Budget has fallen from 26.3 per cent in 1987-88 to 22.8 per cent in 1997-98, while the share of income tax revenue has risen from 64.0 per cent to 71.9 per cent over the same period.

Without comprehensive tax reform, the tax mix will continue to shift automatically towards a greater reliance on direct taxation. Those who can afford expert advice or who (because of the size of their income) have more options open to them to minimise their direct tax will continue to have an advantage over those Australians who cannot afford such advice or do not have tax minimisation options available to them.

In short, under the current tax system, ordinary Australian wage and salary earners will carry an ever greater share of the tax burden.

It is interesting to examine the comments of the then Treasurer on the 1993-94 Budget measures to increase wholesale sales tax rates. In his Budget Speech, he said that `(t)hey will halt the erosion of Commonwealth indirect tax revenue thus preventing Australia's tax system becoming lopsided…Failure to act on the indirect tax side would create pressures for more weight on income tax, a weight which would fall most heavily on middle-income earners'.

The revenue erosion is once more evident, with the indirect tax to GDP ratio presently about one-and-a-half percentage points lower than its level of a decade ago. This tax reform package makes up less than half of that lost ground. But it does so in a way that differs markedly from the 1993-94 Budget strategy. The GST will ensure that the erosion of indirect tax revenue is halted permanently.

The Government notes the report of the Democrats with respect to the current tax system being broken. The Government agrees with the Democrats and without repeating the quotes and their analysis of budget figures, it agrees that a strong case has been made for reform.


The Cameos

The results of Treasury's distributional analysis are presented as a series of cameos, which estimate the impact of the tax package on individuals and families in a variety of situations.

The cameos show the effects of income tax cuts and increases in family benefits and social security payments to be received by various types of household at a variety of private pre-tax income levels. It is assumed for these cameos that the tax package does not affect private pre-tax incomes. Disposable income is affected, however, through the tax cuts and the increased Government cash payments.

The calculations assume that everyone who satisfies the criteria for a Government cash payment does receive that payment. The calculations are based on the projected values of family and social security assistance in July 2000, under pre-tax reform conditions. Then a further calculation was made, again using those projected values but after the application of the tax reform package.

Effects on Consumer Price Index

The change in the CPI represents the increased cost of living produced by the introduction of the tax package. Treasury estimates that this will be 1.9% in 2001-2. However, the calculation does not include the effect of the tax package on tobacco. It does include the effect of the new First Home Owners' Scheme on new house prices.

One of Treasury's assumptions in this area has drawn considerable comment. The calculations have assumed that all households face the same cost of living adjustment as a result of the tax package, i.e. the 1.9%. Most other modellers, while recognising that HES data has flaws, disagree with this approach and prefer to assign different cost of living measures depending on household income levels. Treasury's objection to this method is that it relies on using and manipulating data from the ABS' Household Expenditure Survey in ways which Treasury considers produces unreliable results – this issue is discussed below under the heading – the data debate.

The Treasury further noted that the uniform cost adjustment method simply continued a long established government practice of increasing benefits based on movements in the population-wide CPI.

The Use of Disposable Income

In calculating the cost of living increase, the CPI increase is applied by Treasury to household disposable income, not to consumption. This technique measures the change in real disposable income and avoids the problem of differences in saving ratios between households. It also reveals changes in households' ability to consume or save as a result of cost of living changes.

Treasury argues that avoidance of the problem of different saving ratios does not produce a significant effect on the cameo results – especially since the population-wide household saving rate is only around 5%.

The Treasury Results

Based on the Government's proposed tax package, the Treasury has predicted that prices will rise, on average, by 1.9% in the second year of implementation of the package. This figure includes the effect of the new First Home Owners' Scheme on new house prices and excludes the tobacco prices. [13]

Included in the predicted result is allowance for a 4% increase in pensions and benefits and a promise that over time that increase will be maintained at 1.5% above the CPI impact.

Costs facing Australian exporters are expected to fall by about 3.5% (i.e. by about $4.5 billion a year). Import competing firms will also benefit from reduced costs. The entire package of indirect tax reforms is expected to reduce business costs by 3%. The cost of private investment goods is expected to fall by around 7%.

Rural and regional Australia will also benefit from a fall in transport costs, due to the reduction in diesel fuel excise for off-road use from 43c per litre to zero. For large transport users (including rail) the rate will fall from 43c per litre to 18c per litre. These reductions are in addition to a refund of around 7c per litre which is available to all business users of petrol and diesel. Treasury said there should be no price rise for other consumers.

The Data Debate

The evidence given to the Committee revealed a fundamental difference of opinion between the Treasury and certain commercial/academic modellers on the value of the Household Expenditure Survey (HES).

Two main measures are used in this area: (i) the Australian Bureau of Statistics' (ABS) HES and (ii) the Consumer Price Index (CPI). The HES is carried out every five years by the ABS and surveys in detail the expenditures of some 8400 households across Australia. The difference of opinion between Treasury and certain other modellers, concerns the value to be given to the HES data and its reliability in measuring the effects of some aspects of the tax package.

In evidence given to the Committee, Treasury representatives said that they considered the HES data to be “invalid” for the purposes of distributional analysis.

In response to a statement from Senator Murray that “As soon as you try to disaggregate it (HES data) and get down to small cameos of household numbers, your margin of error soars” [14], Treasury representative, Dr Henry, commented:

Yes … Even at the aggregate level, the household expenditure survey is not valid – or, at least, let me put it this way: if the household expenditure survey is valid, then the Australian national accounts are wildly inaccurate and the CPI is even more wildly inaccurate. [15]

He continued:

… the ABS does not believe the HES. The ABS takes the HES results. It then uses a lot of judgement, a lot of guesswork, a lot of manipulation in order to manipulate those results of the HES to produce the official CPI weights.

… So, even at the aggregate level, the ABS is saying that the household expenditure survey does not provide an accurate representation of household expenditure patterns. [16]

Those modellers who use the HES extensively, however, recognise the flaws in the data set, make adjustments to allow for those flaws and then use it because it has no equivalent available in this country. There was no evidence provided however that conclusively demonstrated that one approach or set of assumptions was better than the other.

Another major concern with the use of HES data is dissaggregation. Treasury have also consistently argued that disaggregating the population-wide CPI among different households is not desirable:

Certain features of the HES invalidate its being used in this way.

This is because of the adjustments made by the ABS in order to calculate weights for the CPI these adjustments assure a better CPI but, without them, the HES data contain biases. For example, adjustments are made for known cases of under reporting (eg tobacco and alcohol) and some expenditure estimates (eg some consumer durable items) are subject to high sampling errors and are adjusted using other data on expenditure levels. [17]

The HES is conducted over a full year to take account of seasonal variations in expenditure. Treasury believe that:

It is not appropriate to draw statistical inferences from individual household records in the survey. It is only through the aggregation process that seasonal and other irregular variations in expenditure patterns are ironed out. [18]

The broader issue of different cost of living measures for different household groups has been examined in detail by the ABS. They concluded, in the publication Feasibility of Constructing Price Indexes for Special Population Groups (Catalogue No. 6445.0) that there was almost no difference in the overall price increase each household group faced over a period of eleven years — despite prices rising by 120 per cent over that period.

For these reasons, governments have consistently adopted the practice of providing compensation for price increases by adjusting benefits by the movement in the population-wide CPI.

As mentioned earlier, and in the ANTS document, the Government has calculated the increase in the cost of living by applying the increase in the population-wide CPI to household disposable income, rather than to household consumption. This approach has been criticised by some witnesses for not calculating the saving ratios among households when determining changes in cost of living. The Government believes that their approach, as stated in the ANTS document:

… is consistent with the approach of many researchers who take the view that current consumption (in a particular week, month, or even year) provides a less reliable measure than disposable income of a household's ability to consume. And it is a household's ability to consume that is really of interest, since that is what is affected by cost of living changes. [19]

The Government also believes that there is a more “pragmatic” reason why cameo calculations are based on disposable income, rather than consumption:

While there is a National Accounts measure of average household saving, there are no reliable data on the saving rates of different types of households. For example, the HES, which has sometimes been used for such purposes, does not contain reliable data of this sort  a point that the ABS emphasises with each survey release.

… Those who would prefer to see cost of living calculations based on current consumption will presumably want to argue that, to the extent households spend less than their total income (ie save), the cost of living increases reported here will be overstated and, hence, the cash gains are understated. They will also take the view that, to the extent households spend more than their income (ie dissave), the cost of living increases will be understated. Making such adjustments, however, would have an imperceptible effect on the cameos especially as the population-wide household saving rate is about 5 per cent. [20]

Adequacy of Compensation

Under the Government's tax reform package people will be overly compensated for the inflationary impact of the introduction of the GST. Personal income tax cuts totaling over $13 billion per year will be introduced from July 2000. Family assistance will be increased by more than $2 billion a year under the new tax system. Tax relief for families is introduced and incentives for low and middle income families are improved in two ways:

families can earn more income before family payments start to be reduced; and

the phase out of payments occurs at a slower rate.

The Government has ensured that all Australians will be able to share in the benefits of tax reform. In particular, families, pensioners and other recipients of income support payments will benefit from increases in their payments.

The Government will increase all pensions and other income support payments (such as the Age Pension) by four per cent which more than offsets the general price rise associated with the GST. The Government will also pay bonuses to eligible senior Australians to help maintain the value of their savings. As the ANTS document states:

These increased rates of assistance raise the maximum level of all income support payments by more than the impact of tax reform on prices (as measured by the CPI), overcompensating recipients for the cost of living effects of the changes to indirect taxation arrangements. The Government will ensure that income support payments are 1.5 per cent higher than they would have been had the normal automatic indexation arrangements applied. [21]

There has been a variety of micro-modelling performed on the adequacy of the compensation provided in ANTS. These include: Treasury's PRISMOD simulation – which was used in ANTS, HES based distributional analysis, and the Melbourne Institute modelling.

As already discussed, the PRISMOD simulation showed that all groups would be better off under the new tax system, with the average family $40-50 per week better off.

The HES data also showed that each household group would be better off, with the cost of living increase varied between 1.3 per cent and 2.5 percent.

The Melbourne Institute modelling is best described in the ACOSS submission as:

(attempting) to take account of variations in household spending and saving patterns by using data from the ABS Household Expenditure Survey (HES) which is adjusted in various ways to improve reliability. Their model also produces a different (and higher) set of estimates of the effects of the consumption tax changes on housing and tobacco prices and on the CPI overall. Their estimate of the average CPI increase is 2.44% compared with Treasury's estimate of 1.9%. [22]

ACOSS asked the Melbourne Institute to model the price effects for households whose primary income source is government income support payments, described by ACOSS as “households with the lowest incomes”. Four assumptions were then modelled, using the Treasury inflation figure of 1.9% as the most optimistic assumption, and then getting progressively more pessimistic (See ACOSS submission p.9.).

The results of this modelling showed that even under the most pessimistic assumption used by the Melbourne Institute, every group modelled was better off under the tax package.


There has been some discussion about the results of the economic modelling commissioned by the Committee indicating that the macroeconomic impacts are very similar for the Government's existing tax package compared to a package which exempts food from a GST and has compensating reductions in income tax cuts.This result is hardly surprising given that the budgetary impact and therefore the fiscal stimulus delivered by both the packages is unchanged. The macro-modelling results therefore provide no support for the position that food should be GST-free.On the equity side, not one of the micro-models presented has been able to show that any individual group will be worse off under the tax reform package, regardless of the inflation assumptions used.

This evidence clearly shows, not only the need for tax reform, but the overall adequacy of the Government's compensation package. For reasons of equity, efficiency and simplicity, the Coalition Senators are strongly against exempting food, or other “necessities of life”, from the tax package.

Further, many of the witnesses who made submissions to the committee targeted the issue of exempting food directly, raising major concerns about any proposal to exempt food. A general summary of most of the concerns raised was provided by Mr Fergus Ryan, Chairman, Business Coalition for Tax Reform.

In summary, excluding food would add significant compliance costs, particularly for small business; tend towards creating a culture of non-compliance and activity that seeks tax avoidance; increase the administrative costs and the time spent by tax administrators and the courts in deciding what is and what is not food for the purpose of tax; and reduce the general competitive advantage of a single rate, broad based indirect tax.

… would reduce the amount of GST paid by the wealthiest 20 per cent of households by double the reduction for poorest 20 per cent of households. In other words, for every dollar of benefit received by the poorest 20 per cent of households, granting a GST-free status for food would deliver a benefit of $7.60 to the rest of the population. [23]

The myth of an Economic Efficiency gain from exempting food

The Senate inquiry has not received any evidence which supports the proposition that there is an economic welfare benefit from exempting food from the GST. In fact the evidence has quite the contrary.

Professor Dixon in his report to the Senate inquiry stated that:

“In the long run, exempting food has a negligible, but negative impact on economic welfare under either labour market assumption. [24]

He also provided this further caveat to his findings:

Our food-exempt simulations bring two issues to mind. The first concerns the costs of implementation, compliance, administration and rent-seeking. These costs are likely to be increased if the GST is implemented with substantial exemptions, but in all of our simulations they have been ignored. These ignored costs should be set against any benefits that we show in our simulations for the tax package, especially in assessing the benefits of exempting food. [25]

The Econtech research also showed that a slight erosion to economic welfare would result if food was made GST-free.

Professor Neil Warren, in his December 1998 paper, explained how the perceived benefits to low income earners, and any suggestion of a benefit to economic welfare, may not be achieved:

There is also another issue why concessional rates for food need to be avoided. This is concern about whether the full benefits of the concession are actually passed onto the consumer. It there are two sectors and one is taxed and the other untaxed, consumers will substitute the untaxed good for the taxed good. As they do this, the price of the untaxed good will rise as a result of increased demand and the price of the taxed good will fall so that not all of the tax burden on the tax good will be passed through to the consumers of that product. As a result, the supposedly untaxed good effectively is taxed because of the increased demand for this good and the pricing response by producers. [26]

Given that the evidence shows that exempting food from the GST has negative impact on general economic welfare, and that there is no evidence that it would provide any specific welfare gains to those on low incomes, there is a strong economic efficiency argument for leaving food in the GST net.

The revenue implications of exempting Food

Significant concerns have been raised by many witnesses about the magnitude of revenue required to remove food from the GST net.

The equity issues surrounding the inclusion of food in the GST net cannot be fully assessed without an understanding of the magnitude of revenue involved. Given the definitional problems surrounding what should constitute taxable food, most witnesses decided to define food widely – excluding restaurants and takeaways.

Treasury estimated [27] that the amount of revenue forgone if food (other than food sold in restaurants and takeaway food establishments) were to be “zero-rated” would be $4465 million in 2000-01, blowing out to $5820 million in 2003-2004.

This evidence is consistent with Professor Warren estimate in his December paper:

“If food was zero-rated, this would reduce revenue by over $5 billion” [28]

The only witness who defined food narrowly for costing purposes was Prof. Dixon who defined it as “… meat, vegetables, bread – all the staples” [29]. Under this narrowest of definitions, he predicted it would cost $2.7 billion a year, by crudely taking 10% of what Australians currently spend on those staples.

Distributional Impact of Making Food GST-Free

Given that most of the proponents for making food GST-free cite equity arguments, it is important to determine who will receive the greatest benefit from so doing.

Professor Warren points out that rather than being a progressive equity measure, zero-rating food would benefit the wealthy to a much greater extent than the poor:

… zero-rating food benefits most in nominal terms, those in the highest income groups. In fact, the top 20% benefit twice as much in terms of dollars and cents from zero-rating food, as does the bottom 20% [30]

Ms Angela Ryan agreed with Prof. Warren's assessment:

If you are looking at the dollar amounts of GST that you are going to lose if you take out something like food, it is the high income earners that will do best out of that system by a proportion of about two to one if you compare the top 20 per cent of households with the bottom 20 per cent. [31]

At worst the distributional impact of exempting food is in fact the opposite to the objectives of those who would wish to do so on progressive equity grounds:

It is reasonable to expect that this action (zero-rating food) would increase the tax burden on non-food goods and services which would provided the greatest benefit to the low-income groups. [32]

Moreover, at best, the distributional impact of exempting food from the tax net shows that it is an extremely blunt instrument for achieving any form of “equity” outcome.

The simple fact is that this approach to compensating for the effects of taxing food is quite inferior to directly targeting those groups which are adversely affected. This should be clear from the magnitude of the benefit to the lower income groups from zero-rating food. In effect, what zero-rating food has done is to reallocate around $2b from a 4.9% GST which was initially on food to all other goods which now have imposed on them a 5.8% GST. [33]

And, Angela Ryan concluded,

… if you include food in the base, you have got more than enough revenue collected from the higher income earners to compensate the lower income earners and have money left over that you can spend on other welfare programs if that is what you want to do. [34]

Conceptual Problems in Defining Food

A common theme in the evidence before the committee was that defining what constitutes food for taxation purposes would lead to massive compliance costs, both to business and the government.

Problems in defining what is food arise because what constitutes food under one classification may not be taxable under another. Which classification has precedence and therefore results in the item being taxable or non-taxable needs to be clear. [35]

In describing the difficulty of defining food, Prof. Warren uses the following examples:

… confectionery is "any item of sweetened prepared food which is normally eaten with the fingers". Clearly, this has a very wide interpretation and not surprisingly, needs to be elaborated upon at length to make it operational. As a consequence, statements such as the following in the UK VAT information on biscuits is typical:

Biscuits covered or partly covered in chocolate or some other product similar in taste and appearance to chocolate are standard-rated. This includes:

(Source: Section 2.5, UK Inland Revenue VAT Notice 701/14 Food, which is reproduced in Attachment C)

A review of the criteria used to define what is and is not food in the VAT legislation applied by various countries results in several common conceptual problems for those drafting the VAT legislation. These problem areas can be categorised along the lines detailed in the following table. [36]


Table 8 Defining Food in Practice under a GST [37]
CriteriaTaxable (and therefore not FOOD)Not Taxable (FOOD)
Temperature Above ambient temperature

Frozen yogurt

At ambient temperature

Refrigerated yogurt

Timing of ConsumptionImmediate consumptionLater consumption
Sweetness Sugar, chocolate or yogurt coveredNo sweet topping
Use of FingersTo be eaten with fingersNot eaten with fingers
Saltiness SaltedUnsalted
Number 1 pudding

1-5 buns

2+ puddings

6 buns (maybe of any type)

Volume 1 pudding less than 425ml

1 beverage less than 500ml

Size of the biscuit ( eg Gingerbread biscuit vs Gingerbread house)

1 pudding greater than 425ml

1 beverage greater than 500ml

Concentration Drink with >25% fruit juice concentrationFruit juice flavoured drink with less than 25% juice
Alcoholic contentDrink with alcohol is taxable if otherwise not taxableNo alcoholic contents

Prof. Warren concludes:

If any one of these criteria could be applied directly to a particular product, then life for those administering the GST legislation is probably not that difficult. However, if a product could be applied to more than one of the above classifications so that under one criteria it is taxable and another not taxable, then very real problems will arise. In this case, the product needs to be uniquely classified - hence the need to endlessly clarify the status of new food products as they are developed and enter the market place or as producers find avenues for making their product non-taxable. [38]

If we take the issue of temperature in isolation, it could have definitional ramifications felt in almost every supermarket, convenience or corner store, as Joycelyn Morton of the ASCPAs pointed out:

… you might have a situation where you have bottles of coke on the shelf which are warm, and there would be no GST on them, but if they were in the fridge and they were cold, they might be deemed to be takeaway and therefore subject to GST. Similarly, in a large supermarket, if they have the large chicken rotisserie, is that takeaway chicken or is that to consume at home? … So you would get all these definitional problems. Every time you try to make that distinction you will be causing people an enormous amount of distress from an administrative point of view. [39]

Administrative and Compliance Costs of a GST

One of the areas which has not been directly modelled by any of the witnesses before the Senate Committee is the effect on compliance costs of excluding food from the tax net. However, evidence has been overwhelmingly supportive of the proposition that exempting food will lead to more rather than less compliance costs. Prof. Warren states in his December paper that:

Clearly, zero-rating or even concessionally treating food has major implications for both the administrative cost of the GST (for the Australian Tax Office) and for the compliance cost of businesses. Together, administrative and compliance costs constitute the operating costs of the tax system. [40]

Compliance costs of the current tax reform package are considered to be easily manageable, and, as Kenneth Claughton of the ASCPAs points out, in the long term beneficial.

… I have a full-time practice at the moment, giving advice, advising clients how to comply with their sales tax obligations. That practice will almost disappear under a GST. [41]

However, the same can not be said if food were to be exempted:

… as long as any GST is kept simple and therefore easy to administer (which in practice means a very broad tax base and a single tax rate), most businesses can use cheap computers and accounting software packages to keep track of their tax liabilities. This may also have the advantage of improving the information flow to managers, an issue that is often cited as a cause of small business failure and a benefit flowing from the introduction of a GST.

If food is zero rated in Australia, then there will be many small businesses who sell food products who will be administering a GST only to claim back GST refunds, thus contributing nothing to GST revenue nor receiving any cash flow advantage. Quite simply, if a multiple (positive) rate GST is introduced, then the burden of compliance could be substantial for small taxpayers (as is noted for the UK below). [42]

Ms Ryan from the ACPAs described excluding food from GST as an administrative nightmare. She went on to say:

… it would be a bonanza for lawyers and, I can say, for Accountants. In lots of ways we are not talking our own book here. If we want to have huge accounting GST practices all around Australia, please exempt food. [43]

The added complexity in compliance from exempting food also leads to higher levels of avoidance, according to the ASCPA

I think it was quite justified that we saw comments from the Labour Party leaders during the campaign that the UK had a real problem with avoidance and evasion. I would contend that the main reason for that the UK has a number of very wide ranging exemptions in their VAT, including food and children's clothing. Under a system like that, it is a relatively easy step for people to be able to falsify invoices. Instead of showing it up on your till as adult clothing you could show it as children's clothing and things like that. This means that the system is not comprehensive. [44]

On the issue of managerial benefits that would accrue to small firms through GST reporting, Prof. Warren wrote:

In the survey of VAT relating to 1986-87, 30 per cent of the respondents agreed that record keeping for VAT gave them benefits. The principal benefit was saving money because respondents did more of their own accounts, giving less work to outside advisers. Other benefits included better stock control and fewer bad debts. [45]

Further he said:

If the administrative and compliance costs of a GST are to be minimised, the GST structure should be simple, with a minimum of rates and exemptions. A simple comprehensive structure avoids difficult borderlines that put up administrative and compliance costs. [46]

To sum up:

All the above discussion clearly highlights the significant downside from an administrative and efficiency perspective, when food is made GST-Free. Even from an equity perspective, while zero-rating food does assist the poor, it is a very ineffective method of compensating those adversely affected by such a tax. [47]


The Government's tax reform plans were comprehensively put before the Australian public in August 1998 in the policy document Tax Reform: not a new tax a new tax system (ANTS).

Implementation of the Government's policy will see:

Personal income tax cuts totalling $13 billion a year providing a 30 per cent marginal tax rate for over 80 per cent of taxpayers

Increased family assistance to low and middle income earners

Increased pensions and government benefits

$10 billion a year of embedded taxes being removed from business

Costs to exporters falling by $4.5 billion.

The Government Senators reach the following conclusions from this stage of Committee's proceedings:

1. On the basis of evidence heard by the Committee and on the basis of the economic modelling commissioned by the Committee, Government Senators conclude that there is substantial support for the Government's tax reform plans as set out in the policy document Tax Reform: not a new tax a new tax system (ANTS).

Economic commentators and modellers confirmed that benefits to the economy as a result of ANTS will come from:

A boost to economic welfare

GDP growth

Employment growth

Export growth

Increased investment

Continued fiscal responsibility; and

More sustainable Commonwealth-State financial relations.

2. There was clear support in the evidence from a wide range of witnesses from the business sector, the unions, farming representatives, academics and tax practitioners for the Government's view that the current tax system is broken. It was accepted that the existing indirect tax revenue base would continue to diminish as the services sector increases its share of the economy.

3. No case was made for making food or other necessities of life GST-free. While macro-modelling showed similar results for the Government's existing tax package compared to a package which exempts food from a GST and has compensating reductions in income tax cuts, the result is hardly surprising given that the budgetary impact and therefore the fiscal stimulus delivered by both the packages is unchanged.

4. No evidence was submitted which showed that the Government's compensation measures were inadequate. The up-front 4 per cent increase in pensions and other government benefits and the guarantee to ensure that income support payments are 1.5 per cent higher than they would have been had the normal automatic indexation arrangements applied, was shown to overcompensate specific groups of low income earners identified by other distributional modellers.

Senator Alan Ferguson Senator Brian Gibson Senator Bill O'Chee

Liberal Party Liberal Party National Party


[1] Evidence, p.20.

[2] Evidence, p.21.

[3] Evidence, p.626.

[4] ECONTECH, submission to the Senate Select Committee on A New Tax System, Modelling a New Tax System (ANTS) – Comparing MONASH and MM303, 14 February 1999, p.20.

[5] Tax Reform: not a new tax, a new tax system (ANTS), p.162.

[6] Evidence, p.116.

[7] Prof. Peter Dixon and Maureen Rimmer, “The Government's Tax Package: Further Analysis based on the MONASH Model”, Centre of Policy Studies, Monash University, Report prepared for the Senate Select Committee on a New Tax System, January 25, 1999, p. ii.

[8] ECONTECH, op cit, p.39.

[9] Ibid, p.19.

[10] Dixon. op cit, p.ii.

[11] David Johnson and Rosanna Scutella, “Long term effects of the governments tax package”, Melbourne Institute of Applied Economic and Social Research, 18 September 1998, p.1-2.

[12] ECONTECH, op cit, p.21.

[13] ANTS, p.162.

[14] Evidence, p.37.

[15] Evidence, p.37.

[16] Evidence, p.38.

[17] ANTS, p.162.

[18] ANTS, p.162.

[19] ANTS, p.163.

[20] ANTS, p.163.

[21] ANTS, p.56.

[22] ACOSS, Submission to the Senate Select Committee on A New Tax System, Modelling the effects of the proposed tax reform package on low income households, December 1998, p.7.

[23] Mr Fergus Ryan, Chairman, Business Coalition for Tax Reform, HANSARD, Senate Select Committee for a new tax system, Wednesday, 3 February, 1999, p. 608

[24] Dixon, op cit, p. iv.

[25] Ibid. p.iv

[26] Neil Warren, “Food: staple of life or staple of the GST?”, ATAX, Faculty of Law, University of NSW, 16 December 1998, p.7

[27] Supplementary question to Treasury on 19 January 1999 from Senator Murray

[28] Warren, op cit. p.7

[29] Evidence, p. 525

[30] Warren, op cit. p.7

[31] Evidence, p. 567-568

[32] Warren, op cit. p.8

[33] Ibid, p.8

[34] Evidence, p.568.

[35] Warren, op cit. p.11

[36] Ibid. p.11-12

[37] Ibid. p.12

[38] Ibid. p.12

[39] Evidence, p. 570.

[40] Warren, op cit. p.13

[41] Evidence, p. 573.

[42] Warren, op cit, p.14

[43] Evidence, p.572

[44] Evidence, p. 565

[45] Warren, op cit. p.15

[46] Ibid. p.16

[47] Ibid. p.17