Reforming State Taxation
By Owen Gabbitas and Damien Eldridge
Click here for PDF version

Assessing options for tax changes at the state level

Recent discussion of tax reform has focused on the Commonwealth tax system. There has beenÊ relatively little systematic discussion of state and territory (hereafter ÔstateÕ) tax arrangements, beyond the need to address the high degree of vertical fiscal imbalance. While these aspects of tax reform are vitally important, there is more to state tax reform than resolving the vexatious question of Commonwealth/state financial relations.

The state tax system is in desperate need of an overhaul. Aspects of it are anachronistic, a legacy of bygone days (e.g. cheque duty and other stamp duties). The legal standing of some taxes is dubious, as highlighted by the August 1997 High Court ruling that franchise fees were unconstitutional. Others taxes areÊ novel attempts to tax selected services within the confines of the Australian Constitution (e.g. bed taxes). Financial deregulation has greatlyÊ increased the range of financial instruments, undermining state financial tax bases, while technological change and globalisation willÊ erode revenue further.

A thorough and com-prehensive overhaul of the state tax system would undoubtedly require the cooperation of the Commonwealth. It may even require constitutional amendment øÊ something that historically has been difficult to achieve. Yet not all reform options are outside the control of state and local governments. There are many worthwhile improvements that the states themselves could undertake independently of any Commonwealth initiatives, either on their own or collectively. The directions for state tax reform identified here generally do not presume any action on the part of the Commonwealth. Nor do we presuppose any change in state government expenditure. Instead, we consider tax reform within a revenue neutral context, recognising that the revenue forgone by abolishing inefficient or inequitable taxes will need to be replaced from other sources.

Overview of state and local government taxation

State and local governments used over 50 different taxes to raise more than $34 billion of tax revenue in 1995-96 (Figure 1). This is equivalent to one dollar in every four raised nationally, or $1 900 per person. States tax a diverse range of activities, from payrolls to gambling, from land to health insurance, and from hire purchase agreements to car parking spaces.

The main sources of state tax revenue are taxes on payrolls, land, financial transactions, motor vehicles and gambling (Figure 1). Franchise fees were also an important source of state tax revenue prior to their being ruled unconstitutional. The interim arrangement designed to
safeguard state tax revenue effectively sees the Commonwealth levying equivalent taxes on behalf of, and then transferring the revenue to, the states (Commonwealth Treasury 1997).

Our analysis includes local government taxes, as local governments derive their limited taxing powers directly from the states.1 Local government taxes also impact on the efficiency and equity of state taxes, particularly where land is subject to both land tax and municipal rates.

Figure 1: Composition of tax revenue, 1995-96.

Assessing the current mix of state taxes

A well-designed tax system would raise the required tax revenue while minimising, as far as practicable, the impact on economic efficiency. To do so, it would generally limit the effect on consumer and producer behaviour. The system would not impose undue compliance costs on taxpayers, or administrative costs on state tax authorities. It would use tax bases that minimise the scope for avoidance or evasion. While income redistribution need not be a prime concern of state governments, state tax systems should avoid exacerbating any inequities. The tax assignment literature indicates that, within a federation such as Australia, the highest tier of government should address equity concerns (e.g. Musgrave 1983).

State governments prefer tax bases that generate revenue that grows in line with essential requirements for public services.

This evaluation covers four key areas of state taxation ø payroll taxes, taxes on land, financial taxes and franchise fees ø addressed in an earlier study by Gabbitas and Eldridge (1998). Although taxes on motor vehicles and gambling raise more revenue than financial taxes, the collective efficiency losses of the latter are likely to be higher. Franchise fees are included as the replacement measures imposed by the Commonwealth effectively continue the previous arrangements.

Efficiency

The most efficient way of raising revenue would be to equate across all taxes the marginal losses in efficiency from raising an extra dollar of revenue (Ramsey 1927). Optimal tax reform would reduce reliance on taxes with high marginal efficiency losses, replacing the revenue forgone from taxes with low marginal efficiency losses. This would allow governments to raise the same amount of revenue while reducing the size of the overall loss in efficiency.

The relative sizes of these losses are an important first step in identifying possible reform options. We therefore estimate the marginal loss in efficiency, or marginal excess burden (MXSB), for a number of key state taxes. The measure is a partial equilibrium measure, in that it only measures the effect in the market subject to the tax, abstracting from wider interactions between taxed activities. An increase in franchise fees on wine, for example, may change the amount of tax revenue obtained from spirits, if consumers switch between alcoholic beverages. This variation in tax revenue may alter the size of the loss in efficiency per dollar of revenue raised. (The degree to which consumers switch between alcoholic beverages is, however, a contentious issue ø Scales, Croser and FreebairnÊ (1995) ).

The MXSB will be larger, the higher is the overall tax rate, and the more responsive demand and supply are to changes in price (i.e. the more elastic the demand). The MXSB of state and local government will be affected by pre-existing Commonwealth taxes, as these raise the overall tax rate.

State tax rates range from as low as 1 per cent on land used for owner-occupied housing (municipal rates), to 100 per cent on tobacco (Table 1). This gives a rough indication of where the efficiency losses are likely to lie. But, as noted, the efficiency of state taxes is also affected by any Commonwealth taxes. In some cases, relatively low rates of state taxation are levied on bases that attract no additional Commonwealth taxation (e.g. land), so the efficiency losses are likely to be very low. In other cases, relatively modest rates of state taxation are imposed on commodities that attract very high Commonwealth taxation (e.g. spirits and petroleum products), so the efficiency losses associated with the state taxes could be much higher than the state rates alone would imply.

The efficiency cost of state taxes also depends on the price responsiveness of demand and supply. Factors such as consumer tastes, the availability of substitute products and taxpayer mobility influence the responsiveness of demand. Changes in resource availability and technological factors, among other things, drive the degree of responsiveness on the supply side. The preferred estimates of the relevant elasticities of demand and supply are also shown in Table 1, where estimates between zero and one in magnitude indicate relatively low price responsiveness, and estimates between one and infinity indicate high price responsiveness. The estimates are taken from more comprehensive surveys of the literature.

Irrespective of the elasticities of demand (the usual bone of contention in debates about optimal taxation), a key feature of the estimates is that two of the state tax bases ø labour and land ø are non-produced and accordingly have relatively low price elasticities of supply (i.e. their supply is relatively less responsive to the price offered for it). This is a key feature determining the relative efficiency of different taxes.

However, because the elasticities are derived from Australia-wide rather than state-specific studies, the MXSBs derived from these elasticities should be interpreted as showing the efficiency impact of all states and territories together making a marginal change in a given tax rate. The MXSBs do not show the costs or benefits from a state acting unilaterally. For mobile tax bases, the elasticities faced by a single state acting alone would generally be larger than those for the nation as a whole. The calculations therefore provide a lower bound on the efficiency effects of acting alone, butÊ the relative rankings across different taxes need not remain the same.

The partial equilibrium estimates of the MXSB of the main state taxes are also presented in Table 1, in the absence of externalities (to be discussed shortly).

The table excludes financial taxes ø primarily FID (financial institutions duty), BAD tax (bank accounts debits tax), marketable securities duty and loan security duty ø for want of the relevant elasticities. However, there are strong reasons for believing that financial taxes are a highly inefficient way of raising revenue, as a number of previous Australian studies have found (e.g. Campbell et al. 1981, PSA 1995, Wallis et al. 1997). Many large financial transactions can be moved between states, if not overseas. The speed with which the Queensland GovernmentÕs 1995 cut in marketable securities duty on share transactions was transmitted nationally supports such a view. Advances in technology and increasing globalisation heighten the prospect for intra- and international mobility, raising the possibility that state financial tax bases will be further eroded in the future.

As expected from the comments above, raising additional revenue through land tax levied in the presence of municipal rates would cause a negligible loss in efficiency. Contrary to widespread belief, the estimates indicate that payroll tax is also a relatively efficient way of raising revenue. The MXSB is lower for those who are currently exempt (up to 9 cents per dollar of revenue) than it is for taxpayers falling just above the tax-free threshold, or those subject to the maximum statutory tax rate (up to 12 cents). The MXSB is substantially lower than this when the accompanying rate of Commonwealth income tax is less than the top marginal rate. For example, the MXSB of payroll tax falls to between 3 and 4 cents in the dollar when the rate of Commonwealth tax is 21á5 per cent. (This does not suggest, however, that the States should consider undermining the redistributive effect of Commonwealth income taxation by raising payroll tax rates on the low-paid.)

Ignoring externalities, the MXSBs associated with franchise fees are higher than the estimates for land tax and payroll tax ø ranging from 15 to 71 cents per dollar of additional revenue raised (Table 1). This arises from the states levying a modest tax on a commodity already subject to extremely high rates of Commonwealth taxation.

In the absence of externalities, these estimates suggest that land tax and, to a lesser extent, payroll tax are relatively efficient state taxes, while those applying to tobacco, petroleum and spirits are relatively inefficient.

Table 1: Marginal excess burden of state taxation in the presence of Commonwealth taxation (without externalising)

However, externalities are costs (or benefits) imposed on others that are not reflected in prices, or otherwise taken into account. As such, they represent efficiency gains (losses) incurred by society thatÊ should be included in the MXSB. To the extent that taxes discourage the activities producing ÔexternalÕ costs, taxes may improve welfare, up to the point where the marginal social benefit from the taxed activity (including the private benefits) equals the marginal social cost (including the private costs). Beyond this level, any increase in the tax rate will decrease economic efficiency. Taxes on tobacco, for example, may improve efficiency by reducing the need for future taxpayers to fund the health care costs of smokers.2 To the extent allowed by existing estimates, our assessment takes these ÔexternalÕ costs into account in calculating the efficiency losses.

The available estimates of the externalities suggest that the largest external costs are associated with the usage of petroleum products. The estimated external costs of road transport use lie in the range $7 billion to $20 billion per year owing to the high costs of road provision, accidents and congestion. The externalities associated with the consumption of alcohol range from $900 million to $6 billion, while the estimates for tobacco range from $500 million to $800 million per year. It is assumed that there are no externalities relevant for land tax, payroll tax and
the consumption of low alcohol beer.

Not all of these externalities are the responsibilities of the states, nor is statewide taxation necessarily the most efficient way to address them. We assess whether statewide taxation is sufficient to cover those externalities that are most likely to be a state responsibility. This requires an allocation of externalities across jurisdictions, and the allocation and valuation of these externalities are contentious. Nevertheless, their omission couldÊ leadÊ to an overestimate of the loss in efficiency. We use the allocations shown in Table 2 to calculate the MXSB in the presence of externalities (see Gabbitas and Eldridge (1998) for more detail).

When externalities are taken into account, the MXSBs for most taxes decline substantially (Table 2).

The taxes fall into three broad groupings. Petroleum franchise fees have low MXSBs, ranging from 4 to 6 cents per dollar of revenue collected. The estimates suggest that the rates of state taxation applying to petroleum products are insufficient to cover the external costs allocated here to state governments. However, there is still a small MXSB because the state taxes exacerbate the distortions from very
high rates of Commonwealth taxation.

Table 2: Marginal excess burden of state taxation in the presence of Commonwealth taxation and externalities.

Table 3: Assessment of main state taxes (as currently implemented)

Conversely, the MXSBs associated with state taxes on spirits and tobacco are still relatively high (58 and 28 cents, respectively, per dollar of revenue raised). The MXSBs associated with the other taxes on alcohol lie between the two (ranging from 12 to 15 cents).

Sensitivity tests indicate the estimate of the MXSB is quite sensitive to the size ofÊ the assumed externality, more so than to the assumed elasticity of demand. The MXSB associated with the taxation of spirits is an exception owing to the high Commonwealth tax rates. The MXSB of tobacco taxes, for example, would fall to 18 cents if the externality were valued at $2 billion, or rise to 30 cents if the externality were valued at $250 million.

Despite using tools of analysis that, in theory, support different tax rates on different commodities, the overall conclusion is that greater uniformity in state tax rates would generally improve economic efficiency, and generate substantial economy-wide gains. The exceptions to greater uniformity generally occur because of externalities. But even where there is a case for raising state tax rates to cover external costs borne at the state level, it is not clear that the total burden on taxpayers could not be reduced by lowering Commonwealth tax rates on the same commodities.

The analysis also shows that efficient tax bases ø land and, to a lesser extent, payrolls ø are being inadequately exploited, because ofÊ low statutory tax rates, or because of a range of rebates and exemptions. The supply of land is more or less fixed and cannot respond to price changes in the way that the supply of produced goods can. Unlike many other activities subject to state taxation, the ownership of land does not also attract Commonwealth taxation, though Commonwealth capital gains tax may apply to the real increase in certain land values when the land is sold. TheÊ supply of labour is also relatively inelastic, so employers can shift most of the payroll tax burden on to employees. The economic incidence of payroll tax is likely to be similar to that of a labour income tax.3

Equity

Many state taxes appear to be fair, as the tax rates increase with the size of the taxable transaction (e.g. land tax, conveyancing duty and debits tax). In addition, many of the tax bases also constitute important components of wealth (e.g. land, financial transactions, labour income).

Despite this apparentÊ fairness, the state taxes considered rate poorly in equity terms, often because of exemptions and other administrative arrangements (such as tax rates expressed in dollar, not percentage, terms). For example, the exemption applying to owner-occupied properties, but not to rental properties, makes land tax unfair. Franchise fees are particularly regressive, as expenditure on smoking and consumption of the more heavily taxed alcoholic beverages (spirits and beer) is proportionately higher among low income earners.

Administration and compliance costs

Overall, state taxes are not particularly expensive to administer. Land tax is the most expensive because the land needs to be valued. But these valuations also form the basis of municipal rates, so the cost of valuing the land should be apportioned between the two. The cost of raising an extra dollar of revenue from either tax is considerably lower.

Nevertheless, considerable scope exists for governments to lower the cost of collecting revenue. Significant cost savings are possible through greater cooperation between states in coordinating their taxes (especially definitions of the tax base), redesigning their taxes and simplifying compliance procedures. Interstate tax competition should not apply to the tax bases as this unnecessarily increases compliance and administration costs. Instead, the states should restrict competition to tax rates. Administrative costÊ savings may also produce wider benefits ø improving efficiency, equity and, in most cases, compliance costs.

Stability

Overall, the state tax bases appear to be relatively stable, although prone to short-term fluctuations. Such variations may cause the states some financial difficulties if they do not make adequate provisions during periods of above average growth.

Overall assessment

An overall assessment of the main state taxes is given in Table 3. It rates the taxes against the four main criteria ø efficiency, equity, stability and administration/compliance costs ø as well as giving separate assessments for ease of avoidance and evasion. It is based on a more detailed assessment against the criteria in Gabbitas and Eldridge (1998).

No one tax performs well against all criteria

Overall, municipal rates and payroll tax rate well against three of the four main criteria, but poorly against either administration or compliance costs. Land tax, as currently implemented, rates poorly against both administration costs and equity, but could be easily modified to perform well on equity by extending the tax base to include owner-occupied housing. Despite having high compliance costs, payroll tax is one of the cheapestÊ state taxes to administer, because of the relatively small number of taxpayers and large amount of revenue raised. Municipal rates have low compliance costs, although administration costs are very high because of the cost of valuing the land. Nevertheless, additional revenue could be raised at a low cost. Further, these taxes have broad bases, capable of raising substantial revenue.

At the other end of the spectrum, a number of state taxes ø most notably BAD tax, most stamp duties including conveyancing duty, and the franchise fee on spirits ø perform poorly against the key equity and efficiency criteria. In addition, a number of stamp duties raise only modest amounts of revenue. Financial taxes are likely to be particularly inefficient because the tax bases are highly mobile between states and, increasingly, between countries, and many substitute instruments are taxed differently. FID rates better than the other financial taxes on equity grounds, but worse against compliance costs owing to the breadth of its tax base.

The remaining state taxes ø primarily franchise fees on beer, wine, tobacco and petroleum products ø lie in between, performing better against some criteria than others. These generally perform well on efficiency grounds, though poorly on equity grounds. However, externalities associated with the consumption of these commodities argue for keeping these taxes, despite their inequities.

Despite the potential for conflict between the equity and efficiency criteria, the assessment highlights the fact that, in judging state taxes, these criteria tend to reinforce each other. Efficient state taxes also tend to be equitable state taxes, while inefficientÊ ones are generally inequitable. This suggests that the states could raise the same revenue more efficiently and fairly than they currently do.

Assessing various reform options

The states could rectify aspects of their tax system by improving the design and implementation of existing taxes. However, where the efficiency costs of current taxes are relatively large, significant improvement may require lowering tax rates and recovering the revenue elsewhere. This would imply a change in the mix of taxes used. Further improvements could be achieved by extending the scope of state taxation beyond the bases currently in use. However, this would require the assistance of the Commonwealth, and/or amendments to the Australian Constitution.

Improving existing state taxes

The states and local government could improve their tax systems by:

¥ÊÊ making greater use of user charges for services, such as water and garbage disposal, so that users have important information about the costs of providing these services;

¥ harmonising tax bases (i.e. employing standard definitions and thresholds) across states to reduce the incentives for firms to rearrange their affairs across states and to lower compliance costs for firms operating in more than one jurisdiction;

¥Ê reviewing those taxes designed to correct for externalities as well as raise revenue (probably best done in cooperation with the Commonwealth); and

¥Ê addressing any equity concerns through well-designed concessional arrangements and, wherever possible, specifying state tax rates in percentage or ad valorem terms.

Within the current broad tax mix, the states could also improve efficiency somewhat by replacing all state financial taxes with a single broad-based financial tax. Such a tax might resemble FID ø levied on a broad base at a single ad valorem rate ø without a cap on the maximum amount payable. The states could levy the new tax either on deposits (as is currently the case with FID) or withdrawals. In the long term, the two approaches would be more or less equivalent (with some timing differences in revenue collection).

The states could reduce the frequency of monthly payroll tax payments to reduce the high compliance costs associated with the tax. Business would still be required to pay the same amount of tax, but on a less frequent basis. In addition:

¥Ê Queensland and the Northern Territory could expand their payroll tax bases to include employer superannuation contributions ø an important first step in standardising the definition of payroll among states; and

¥Ê Western Australia and the Northern Territory could consider simplifying their complicated deduction schemes, either by moving to a single marginal rate scheme (as in New South Wales) or by employing a simpler deduction scheme (as in Queensland).

With harmonisation, each state eventually should employ the same payroll tax structure (though not necessarily have the same payroll tax rates).

Changing the tax mix

Other ways by which the states could improve the performance of their tax systems would be to change the way certain taxes operate and to alter the mix of taxes used to raise revenue.

While the reform noted above could improve the efficiency of financial taxes to some extent, they would remain relatively inefficient for two main reasons. The tax base ø the size of the financial transaction ø would remain a poor proxy for the underlying service being rendered. And a broad-based financial tax would still cascade along the production chain, continuing the Ôtaxes-on-taxesÕ problem. Its efficiency is also likely to be reduced by technological developments ø such as electronic commerce ø that will dramatically increase the geographic mobility of financial transactions. Consequently, a better option may be for the states to abolish financial transactions taxes altogether, and to raise the forgone revenue by other means. The Coalition governmentÕs tax package would also allow the states to abolish financial taxes (Costello 1998).

Conveyancing duty discourages mobility and is indiscriminate in whom it affects. Although the rate of duty payable increases with the value of the property, conveyancing duty is inequitable in that it applies only to those who move (unlike municipal rates or land tax). When duty is payable, the amount paid is substantial ø both in absolute terms and as a proportion of the underlying value of the transaction ø and affects behaviour significantly. Thus, conveyancing duty is both inefficient and inequitable. One option would be to abolish conveyancing duty and raise the revenue forgone through an increase in land tax.

There appears to be considerable scope for the states to place greater reliance on land tax as a source of revenue. Extending land tax to owner-occupied housing, as New South Wales has done recently, would ensure more equitable treatment of home owners and renters. Such a move would improve both the efficiency and fairness of the land tax.

Broadening the land tax base may cause financial difficulty to low income home owners. If this is the case, the states could consider raising the tax-free threshold. The threshold could be indexed to eliminate the effect of bracket creep brought about by increases in nominal property values. In addition, the states could continue to offer concessional arrangements to those in genuine need (e.g. pensioners).

In its current form, payroll tax is one of the broadest and more efficient taxes used by the states. Thus, it is also a candidate to be used to recover revenue forgone by abolishing relatively inefficient taxes.

The efficiency cost estimates suggest that base-broadening measures would be preferable to raising payroll tax rates. Currently, only 8 per cent of private sector firms pay payroll tax (ABS 6348.0). The current tax-free thresholds cannot be justified on the grounds that the revenue forgone is fully offset by avoided administration and compliance costs. Some form of threshold may be justified on these grounds, but it would be lower than current thresholds. The efficiency cost estimates suggest that payroll taxes could even be raised slightly to replace revenue forgone on other taxes, while still allowing an improvement in overall efficiency.

Once plausible estimates of the externalities associated with petroleum products, alcohol and tobacco use are taken into account, the efficiency costs of state taxes on tobacco and spirits appear relatively high, while those on petroleum products appear relatively low. Hence, the states could improve overall economic efficiency substantially by lowering their franchise fees on tobacco and spirits, and recovering the forgone revenue by raising state taxes on petroleum products. Retail prices of petroleum products need not necessarily increase as there is scope for an offsetting reduction in Commonwealth taxes.

Going beyond current state tax bases

Broadening the current set of state taxes would offer scope to use taxes that are not only more efficient, but also more equitable. However, options that involve a broad expenditure or income base ø allowing a reduction or replacement of the more distorting existing taxes ø would require the cooperation of the Commonwealth and/or amendments to the Constitution.

References

Australian Bureau of Statistics (ABS) 5506.0, Taxation Revenue, Australia, AGPS, Canberra.

Australian Bureau of Statistics (ABS) 6348.0, Labour Costs Australia, AGPS, Canberra.

Campbell, J.K., R.G. McCrossin, A.W. Coates, J.S. Mallyon, K.W. Halkerston and F. Argy 1981, Australian Financial System: Final Report of the Committee of Inquiry, AGPS, Canberra.

Commonwealth Treasury 1997, Constitutional Invalidation of State Business Franchise Fees: Temporary Commonwealth Safety Net Arrangements, Press Release No. 85, 6 August.

Costello, P. 1998, Tax Reform: Not a New Tax, a New Tax System: The Howard GovernmentÕs Plan for a New Tax System, Commonwealth Treasury, Canberra.

Findlay, C.C. and R.L. Jones 1982, ÔThe marginal cost of Australian income taxationÕ, Economic Record, 58:253ø262.

Gabbitas, O. and D. Eldridge 1998, Directions for State Tax Reform, Productivity Commission StaffÊ Research Paper, Ausinfo, Canberra, May.

Musgrave, R.A. 1983, ÔWho should tax, where, and what?Õ, in C.E. McLure Jr (ed.), Tax Assignment in Federal Countries, ANU Press, Canberra.

Prices Surveillance Authority 1995, Inquiry into Fees and Charges imposed on Retail Accounts by Banks and Other Financial Institutions and by Retailers on EFTPOS Transactions, PSA, Melbourne.

Ramsey, F.P. 1927, ÔA contribution to the theory of taxationÕ, Economic Journal, 13:277ø297.

Scales, W.I., B.J. Croser and J.W. Freebairn 1995, Winegrape and Wine Industry in Australia, A Report by the Committee of Inquiry into the Winegrape and Wine Industry, AGPS, Canberra, 30 June.

Wallis, S., B. Beerworth, J. Carmichael, I. Harper, L. Nicholls and G. Smith 1997, Financial System Inquiry, Final report, AGPS, Canberra.

About the Author
Owen Gabbitas is Assistant Director of the Trade and Economic Studies Branch at the Productivity Commission.Ê Damien Eldridge is Assistant Manager of the Emissions Treaty Team, Australian Greenhouse Office. This article is an abridged version of a Productivity Commission Êresearch paper written when both authors worked for the Commission. The views expressed are those of the authors and do not necessarily represent those of the Productivity Commission.

Policy is the quarterly review of The Centre for Independent Studies. For more information on subscribing to Policy, click HERE

If you are interested in the Centre's activities and publications, why not subscribe to e-PreCIS, our regular email update on the latest news and events.

(e-PreCIS requires html capable email facilities, such as Microsoft Outlook Express or Netscape Messenger)