Go back to John Quiggin's home page
Go back to John Quiggin's 1997 articles
Estimating the benefits of Hilmer and related reforms
Professor of Economics
James Cook University
Published as: Quiggin, J. (1997), 'Estimating the benefits of Hilmer and related reforms', Australian Economic Review 30(3), 256-72.
I thank Peter Dixon, Pat Ranald, Nancy Wallace and an anonymous referee for helpful comments and criticism.
This paper shows that the benefits of microeconomic reforms, including the 'Hilmer reforms' have been overestimated in previous studies.
The Industry Commission's 1995 study of the growth and revenue implications of Hilmer and related reforms has been influential in the debate over microeconomic reform. In addition to reform of Telstra, Australia Post, the Federal Airports Corporation, the Civil Aviation Authority, state electricity, gas and water authorities, and rail, road and port authorities, the Commission examined a broad-ranging program including competitive tendering for the provision of public services, deregulation of the building industry and a move to self-regulation of industries such as meat processing. The Commission estimated that the implementation of these reforms would result in a 5.5 per cent increase in GDP. In this paper, the Commission's analysis is subjected to a detailed critique. It is argued that most of the estimated productivity gains are overoptimistic, representing upper bounds to possible achievement rather than likely outcomes. Furthermore, it is argued that the dominant flow-on effects of microeconomic reform will be negative, arising from the fact that at least some of the workers directly displaced by reform will be permanently displaced from the employed labour force.
Estimating the benefits of Hilmer and related reforms
In 1992, an Independent Committee of Inquiry , chaired by Professor Fred Hilmer, was established to examine National Competition Policy. The Committee presented its report in 1993 (Hilmer, Rayner and Taperell 1993, hereafter called the Hilmer Report). Since the release of the Report, the term 'Hilmer and related reforms', or, more simply 'Hilmer', has been used as a shorthand way of referring to that stage of Australian microeconomic reform which focusses on competition policy.
To encourage acceptance of the Hilmer reforms by the states through the Council of Australian Governments (COAG) process, the Keating government requested the Industry Commission to analyse the effects of 'Hilmer and related reforms' on economic growth and on the revenue of state and federal governments. The resulting report, entitled The Growth and Revenue Implications of Hilmer and Related Reforms (Industry Commission 1995a), took a very broad view of the scope of the Hilmer reforms. With the exception of tariff policy, the term 'Hilmer and related reforms' was taken to include the entire microeconomic reform agenda long advocated by the Commission.
In addition to reform of Telstra, Australia Post, the Federal Airports Corporation, the Civil Aviation Authority, state electricity, gas and water authorities, and rail, road and port authorities, the Industry Commission (1995a) examined a broad-ranging program including competitive tendering for the provision of public services, deregulation of the building industry and a move to self-regulation of industries such as meat processing. The Commission estimated that the implementation of these reforms would result in a 5.5 per cent increase in GDP.
The modelling framework used by the Industry Commission (1995a) was based on that used in the Commission's 1989 Annual Report (Industry Commission 1989) where the gains from a fairly similar program of microeconomic reform were estimated at 6 per cent of GDP. However, the assumptions underlying the estimates are presented more clearly in Industry Commission (1995a). Other estimates of the benefits of microeconomic reform, such as those of the Business Council of Australia (1994) and Bureau of Industry Economics (1994), have taken the Industry Commission (1989) as a starting point, but have incorporated additional assumed benefits, yielding claims that microeconomic reform could raise the level of GDP by as much as 20 per cent over a period of 5 to 10 years.
The Industry Commission's (1995a) estimates turn on two classes of assumptions. First there are detailed assumptions about the productivity effects of microeconomic reform in those sectors of the economy that are directly affected, which are referred to as the 'direct' effects of reform. Second, there are assumptions about the flow-on effects to the economy as a whole.
In this paper, both classes of assumptions are subject to a detailed critique. It is argued that most of the estimated productivity gains are overoptimistic, representing upper bounds to possible achievement rather than likely outcomes. Furthermore, it is argued that the dominant flow-on effects of microeconomic reform will be negative, arising from the fact that at least some of the workers directly displaced by reform will be permanently displaced from the employed labour force.
1 Methods of estimating the benefits of microeconomic reform
The Industry Commission (1995a) uses a two-stage procedure to estimate the benefits of microeconomic reform. First, for each sector where reform is expected to take place, estimates are made of likely changes in labour and capital productivity, output prices and other parameters. Second, the ORANI model is used to estimate the change in the general equilibrium levels of GDP, consumption and other variables arising from these changes in parameters relative to a base case where no reform takes place.
The estimated increase in GDP may therefore be divided into two parts. The first part is the direct benefit arising from the productivity gains projected by the Industry Commission. This is the short-run increase in output that would take place if the estimated productivity gains were realised without any adjustment in the labour and capital employed in different industries. Aggregating across the sectors considered in the report, the Industry Commission (1995a) estimates a total direct benefit equal to 2.3 per cent of GDP from the implementation of microeconomic reform (Table 1).
Table 1 Estimated benefits of Hilmer and related reforms (per cent of GDP)
In addition to direct benefits, there is the increase in the level of GDP arising from the adjustment to a new general equilibrium. This second-round benefit is estimated at around 3.1 per cent. Combining the direct and second-round benefits of reform yields an estimated final benefit of 5.5 per cent of GDP, or around $25 billion (Table 1). Assuming that the benefits are expected to be realised over a period of five years, this would correspond to an increase of approximately 1.1 percentage points in the annual rate of growth of GDP.
1.1 World best practice and government business enterprises
An important part of the Hilmer reform program is the creation of a more competitive environment for government business enterprises, notably including telecommunications and electricity enterprises. The Industry Commission (1995a) assumes that exposure to competition will force Australian government business enterprises to achieve 'world best practice' through internal restructuring. Hence the benefits of microeconomic reform may be estimated by measuring the gap between the current performance of these enterprises and world best practice. Drawing on work undertaken by the Bureau of Industry Economics (1994), the Commission identifies a 'benchmark' enterprise and assumes, without specifying what changes in practices need to take place, that reform will induce the Australian government business enterprises in question to emulate the performance of the benchmark enterprise. The critical issue for the world best practice approach is, therefore, the choice of an appropriate benchmark.
World best practice analyses used by the Industry Commission (1995a) typically begin with the presentation of a set of enterprises for which some productivity measure has been estimated. As an initial step, the Commission compares the performance of the Australian enterprise with that of the top-ranked enterprise in the set studied. In most cases, this comparison is invalid because the enterprises are not comparable. For example, labour productivity (measured in gigawatts per employee) in the Australian electricity industry is compared by the Commission to that of a Canadian electricity company serving a few large industrial customers. The small number of customers served by the Canadian company greatly reduces the need for line workers and service technicians, and hence results in a higher level of labour productivity than can be achieved by Australian enterprises serving a large number of customers. After discussing the comparison between Australian performance and that of the top-ranked enterprise, the Commission normally sets the world best practice benchmark at or above the measured performance of the second-ranked enterprise. Even though the resulting benchmark is above any plausible upper bound, it is made to appear more reasonable by reference to the original, clearly invalid comparison.
In any sample of productivity measures for selected enterprises there will be errors, measurement biases and enterprise-specific factors. As an example of a measurement bias, the figures used by the Industry Commission (1995a) fail to take proper account of contracting out. Firms that contract out a large proportion of their work will have spuriously high labour productivity because the employees of contractors are not taken into account in computing the labour input. As an example of an enterprise-specific factor, Sweden has a large electronics industry specialising in the production of telecommunications equipment. This has obvious implications for the supply of skilled telecommunications technicians and engineers, and therefore for labour productivity in the Swedish telecommunications enterprise used as a benchmark by the Commission. The combination of errors, measurement biases and enterprise-specific factors ensures that, if a large enough sample is collected, it must include some firms or countries with exceptionally good measured performance. There are no grounds for supposing that these performances can be emulated by all the firms or countries in the sample.
The difficulties of applying the world best practice approach may be analysed in terms of frontier production functions (Schmidt 1976). In the standard approach, based on linear programming, the production frontier is derived as the convex hull of the observed set of input and output measurements. The production frontier is then constructed by computing the maximum outputs obtainable from linear combinations of these input bundles. The Industry Commission (1995a) fails on several occasions to take proper account of differences in factor input mix. However, even when correctly applied, the linear programming version of the frontier approach takes no account of the effects of errors, measurement biases, stochastic shocks and enterprise-specific factors, so that these effects are confounded with deviations from world best practice .
One approach to overcoming these problems is the use of a stochastic frontier production model (Aigner, Lovell and Schmidt 1977). The key idea is to partition the residual from an estimated production function (that is, the observed differences in total factor productivity) into two components. One component, symmetrically distributed about zero, reflects measurement errors, enterprise-specific factors, and so on. The other component, with a distribution truncated at zero, reflects deviations from best practice. In practice, these two components cannot be distinguished statistically except on the basis of somewhat arbitrary assumptions about the precise distribution of the error terms.
Despite the difficulties of econometric estimation, an example will show the implications of the stochastic frontier approach. Suppose half the variance in measured total factor productivity for the sample as a whole is due to deviations from best practice and the other half is due to measurement errors and other enterprise-specific factors. Then enterprises that adhere to best practice will always be in the top half of the sample. Broadly speaking, the 'average' best practice enterprise will have a productivity level around the 25th percentile for the sample as a whole.
If world best practice benchmarks are to be used to predict the outcomes of successful reform, rather than as a basis for exhortations to increased effort, the use of the 25th percentile productivity level for an appropriate sample would be preferable to the current practice of choosing a maximal or near-maximal observation. Unfortunately, the data published by the Bureau of Industry Economics in its benchmarking studies is not sufficiently detailed to permit derivation of 25th percentile estimates.
1.2 Productivity gains and the choice of counterfactual
In assessing the effects of microeconomic reform on the productivity of government business enterprises, it is necessary to choose a counterfactual, that is, an estimate of the productivity level that would be realised in the absence of reform. The benefit arising from microeconomic reform is equal to the difference between the productivity level that could be achieved under reform and the productivity level that would be achieved under the counterfactual.
Most of the government business enterprises examined by the Industry Commission (1995a) have experienced steady growth in total factor productivity in the past. For example, Telstra has experienced total factor productivity growth rates of 4 to 5 per cent per year over the past twenty years (Arena, Bahtesevanoglou and Branton 1992). However, the Industry Commission treats all productivity growth taking place after reform as being a benefit of reform. This procedure is equivalent to the counterfactual assumption that, in the absence of reform, productivity growth will cease.
To illustrate the effect of the Commission's procedure, consider the following example. Suppose that reforms are being considered over a five year period, that the annual rate of productivity growth observed in the past is 4 per cent and that the Commission estimates that, under reform, the annual rate of productivity growth will increase to 5 per cent. The Industry Commission treats as a benefit of reform all the productivity growth predicted for the five year period -- a cumulative increase in productivity of 27 per cent. The correct procedure is to treat as a benefit of reform only the additional productivity growth of 1 per cent per year -- a cumulative increase in productivity of 5 per cent.
It would seem reasonable to suppose that, in the absence of reform, these rates of productivity growth would be unchanged. However, in view of the general decline in rates of productivity growth since 1973, it might be argued that microeconomic reform is needed, not only to achieve improvements in the rate of productivity growth, but to prevent a deterioration that would otherwise take place.
1.3 Other issues
The Commission's practice of choosing upper bound estimates of the direct benefits of reform is not confined to the inappropriate choice of world best practice benchmarks and productivity counterfactuals, but is a consistent feature of the report. Informal estimates of efficiency gains of 'up to 50 per cent' in the provision of legal services are converted into assumed gains of 50 per cent (Industry Commission 1995a, p. 112). Reductions in the costs of administering and complying with building regulations reported by one local authority are converted into the complete elimination of such costs throughout the economy (Industry Commission 1995a, p. 135). And so on.
In a number of areas, cost reductions arising from lower wages or increased work intensity are treated as productivity gains rather than as income transfers. This problem is most evident in the discussion of competitive tendering and contracting (Section 2.6 ), but it is also important in the discussion of government business enterprises.
1.4 ORANI closure
Decisions made by the Industry Commission (1995a) regarding the macroeconomic closure assumptions for the ORANI model have the effect of substantially increasing the estimated benefits of microeconomic reform. Since all workers displaced by reform are assumed to be re-employed elsewhere in the economy, output must increase. The internal balance condition for equilibrium requires
Y = C + I + G + (X - M)
Y is aggregate output;
C is private consumption;
I is private investment;
G is government expenditure;
X is exports; and
M is imports.
The most important of the closure assumptions, described by the Industry Commission (1995a, p. 50) as 'highly artificial', is that tax rates and government expenditure do not change. Although the reforms are projected to generate increases in government revenue, the higher revenue is not used to finance increased government expenditure G, nor is it returned to the community through tax cuts, which would raise disposable income and therefore increase C. Since the equilibrium investment level I is proportional to capital stock, internal balance can be restored only through an improvement in the trade balance X - M.
The ORANI assumption that elasticities of supply and demand for minerals are very high ensures that most of this adjustment takes the form of an increase in the output of the mining sector. The expansion of the mining sector arising from the Commission's choice of closure assumptions is amplified by specific assumptions regarding the price effects of reform.
Even where microeconomic reform has no direct effect on the mining sector, increases in the output of the mining sector make up a large part of the benefits estimated by the Industry Commission. For example, the Commission projects that removing newsagents' local monopoly on suburban newspaper delivery would increase the level of GDP by 0.03 per cent. Mining output is projected to increase by 0.15 per cent. Since mining contributes a little less than 10 per cent of GDP, the increase in mining output is equivalent to a little less than 0.15 per cent of GDP, or nearly half the total increase in GDP projected to arise from reform of newspaper delivery.
Because the traded goods sector, and particularly the mining sector, is highly capital-intensive, the expansion of the traded goods sector required to restore internal balance implies an expansion of the capital stock, by around 6 per cent, and a similar increase in the capital-labour ratio. A comparison between Columns 1 and 3 of Table 1 shows that the second-round increase in output and productivity estimated to arise from the restoration of internal balance (measured by the difference between Column 3 and Column 1) is larger than the estimated direct benefit of reform reported in Column 1.
The fact that the GDP gains projected by the Commission depend on a substantial expansion in the capital-intensity of the economy means that GDP is inappropriate as a measure of social welfare (Forsyth 1992). The extra growth in capital stock must come either from increased savings in the adjustment period, or from increased foreign debt. In either case, the increase in GDP overstates the welfare gain.
The Industry Commission (1995a) makes no attempt integrate the unemployment associated with microeconomic reform into the medium term ORANI analysis, even though available evidence suggests that many employees who are retrenched or who accept redundancy packages as part of the adjustment to microeconomic reform will not work again for many years. Even where workers are re-employed, it is frequently in jobs requiring lower skills, implying a semi-permanent shift in the effective supply of labour. The medium term economic impact of voluntary withdrawal from the labour force, financed by a redundancy package, is identical to that of an increase in unemployment.
Some proportion of the unemployment generated by labour-shedding during the implementation of microeconomic reform should be modelled as a medium term reduction in the labour force. Evidence reported by the Industry Commission (1995b) suggests that about 50 per cent of workers made redundant by microeconomic reform were still unemployed or not in the labour force after three years (around the midpoint of the ORANI 'medium term'). A 'back-of-the-envelope' approach, based on the assumptions that 35 per cent of workers made redundant by microeconomic reform withdraw from the labour force in the medium term and that the labour share of output is 70 per cent, implies that around 25 per cent (ª 0.7*0.35) of total gains from pure labour productivity increases will be lost as a result of labour force withdrawal.
In Column 4 of Table 1, the adjusted estimates of the direct benefits of microeconomic reform, presented in Column 2, are modified to take account of unemployment effects. It is assumed that 25 per cent of any gain arising from reduced employment, and hence higher labour productivity, is dissipated in higher unemployment or exits from the labour force. As a result of this modification, the estimated total benefit arising from the implementation of microeconomic reform is 0.67 per cent of GDP.
2 Specific areas of reform
In this section, the Industry Commission's (1995a) estimates of the direct benefits of microeconomic reform, presented in Table 1, Column 1 are critically analysed. The Commission analyses the effects of reform in nineteen areas of the economy. For simplicity, the nineteen sectors examined by the Commission are not considered separately here. Rather those sectors in which the issues involved are similar are considered together. The discussion of each sector begins with a brief summary of the reforms under consideration and the direct benefits of these reforms, as estimated by the Commission. The Commission's analysis is then scrutinised, and alternative estimates of the direct benefits, summarised in Column 2 of Table 1, are derived.
The reforms considered for the telecommunications sector are the introduction of Optus as a competitor for Telstra in 1992, and the proposed abolition of the existing legislated duopoly from 1997. The Industry Commission (1995a) projects that a 45 per cent improvement in labour productivity and a 22 per cent increase in capital productivity will arise as a result of these reforms. This implies a direct benefit equal to 0.5 per cent of GDP (Table 1, Column 1).
The Commission's analysis of the effects of microeconomic reform in the telecommunications sector provides a good illustration of the errors discussed in Sections 1.1 and 1.2. First, overoptimistic claims are made about the potential benefits of the achievement of world best practice. The Commission assumes that Australia can equal the labour productivity of Sweden, as measured by main lines per employee and minutes of international telephone traffic per employee. Apart from differences in population density, the main population centres of Sweden are within 1500 kilometres of major European capitals from Moscow to London, making comparisons with Australian costs for international calls invalid. More plausible benchmarks may be derived from comparisons between Australia and the United States or Canada, both of which have average prices about 10 per cent below those in Australia, implying that productivity levels must be around 10 per cent higher.
Second, productivity gains and price reductions in the period since 1992 are treated as benefits of reform. In fact, the overall rate of reduction in the price of telecommunications services has been determined not by competition but by the price cap imposed on Telstra. The price cap merely ensures that the trend rate of decline observed in the price of telecommunications services before 1992 is maintained. The move from monopoly to duopoly has not accelerated the rate of price reductions or of technological progress.
There is no guarantee that the more competitive system of telecommunications regulation proposed for 1997 will generate efficiency gains. Nevertheless, the alternative estimates presented here are based on the assumption that competition will lead to a reduction of 10 per cent in labour and capital costs, yielding a direct benefit equal to 0.2 per cent of GDP (Table 1, Column 2).
2.2 Australia Post
The reform considered for Australia Post is the removal of its monopoly on letter delivery. The Industry Commission (1995a) projects a improvement of 12 per cent in labour productivity, which implies a direct benefit equal to 0.04 per cent of GDP (Table 1, Column 1).
The Commission's estimate reflects an inappropriate choice of counterfactual. Australia Post currently projects that labour productivity will grow at 2.5 per cent per year. The Commission suggests that, under reform, a rate of 3 per cent can be achieved, and treats all of this growth as a benefit of reform. The correct procedure is to estimate the extra labour productivity growth attributable to reform, which is equal, on the Commission's estimates, to 0.5 per cent per year, or 2 per cent over the four-year timeframe considered here. This implies an alternative estimated direct benefit of less than 0.01 per cent of GDP (Table 1, Column 2).
2.3 Airports and aviation
The only Hilmer-specific reform considered for Federal authorities regulating air transport is the imposition of commercial pricing policies. The Industry Commission (1995a) however, argues that a more competitive environment will generate a 12 per cent cost reduction for the Federal Airports Corporation and a 16 per cent cost reduction for the Civil Aviation Authority. This implies a direct benefit equal to 0.03 per cent of GDP (Table 1, Column 1).
The approach used in deriving estimates of the direct benefits of reforms to the Federal Airports Corporation is a variant on the world best practice method. The analysis is based on the assumption that all Australian airports will reach the capital productivity level of the Sydney airport, yielding an 8 per cent reduction in total costs. The Commission takes no account of the fact that the activities of airports are subject to scale economies. That is, the larger the number of flights arriving at a given airport, the lower the cost per flight. For example, the data presented by the Commission shows that Sydney, Australia's busiest airport, was the lowest-cost.
The Industry Commission (1995a, p. 175) claims that 'competition between airports [will] move all airports towards best practice, and result in large labour and capital productivity improvements'. This claim implies that the services provided by an airport in Sydney are a substitute for the services provided by an airport in Melbourne, which is absurd.
For the Civil Aviation Authority , the Commission extrapolates previous productivity gains and treats the whole gain as a benefit arising from the Hilmer reforms. It seems more reasonable to assume that some of the observed gains are attributable to general technological improvements and to scale economies associated with increased output. In view of recent investigations into regulatory failures by the Authority, which may have contributed to fatal air accidents, it also appears that pressure for cost reductions and a more commercial approach have led to a decline in output quality that is not measured in the data used by the Commission.
No evidence is presented by the Commission to show that any net benefits will arise from the Hilmer reforms in the air transport sector. Nevertheless, to avoid an unduly conservative approach, the alternative estimates presented here are based on efficiency gains of 2 per cent for the Federal Airports Corporation and 8 per cent for the Civil Aviation Authority , leading to an alternative direct benefit equal to 0.01 per cent of GDP (Table 1, Column 2).
2.4 Electricity, gas and water
The main reform considered for the electricity and gas sector is the establishment of a national market, allowing interstate competition between suppliers, and removing restrictions on competition between electricity and gas. Pricing reforms and improvements in operating efficiency for water authorities are also considered. The Industry Commission (1995a) projects a labour productivity improvement of 50 per cent for the electricity sector and 15 per cent for the water sector, and a capital productivity improvement of 4 per cent for the electricity sector and 7 per cent for the water sector, as well as a large increase in the proportion of electricity generated using gas. These productivity gains imply a direct benefit equal to 0.6 per cent of GDP (Table 1, Column 1).
The Commission's estimates suffer from most of the deficiencies already addressed in the case of telecommunications. The implied counterfactual is zero productivity growth, but the electricity industry has experienced steady growth in total factor productivity, at annual rates of 3 to 4 per cent, for many years. As the Commission observes, the rate of total factor productivity growth may even have decelerated since microeconomic reform began in earnest in the late 1980s. For example, the Bureau of Industry Economics (1994) estimated a total factor productivity growth rate of only 1 per cent for the period 1989-90 to 1991-92. The Economic Planning Advisory Council (1994), however, found a slight acceleration, estimating that the rate of total factor productivity growth was 4.2 per cent over the period 1989-90 to 1992-93, compared to 3.1 per cent for the preceding decade.
The evidence presented by the Commission gives no grounds for supposing that microeconomic reform has thus far yielded any net improvement in the rate of total factor productivity in the electricity industry. Nevertheless, labour productivity improvements of 50 per cent are projected, based on the world best practice approach. The application of world best practice methods in this case is invalid, since they are applied to a single factor of production, namely labour.
The different techniques for generating electricity vary enormously in labour intensity. Coal-fired stations are the most labour-intensive, followed by gas, nuclear and hydroelectric generation. The greater labour costs of coal-fired stations are offset by higher fuel costs in the case of gas-fired stations and by higher capital costs in the case of nuclear and hydroelectric generation. Yet the Commission uses world best practice measures of labour productivity in which the Australian industry, which is almost exclusively based on coal, is compared with firms which have large components of hydroelectric and nuclear power. In addition, firms that supply a few large industrial consumers are included, even though there are obvious cost savings that cannot be emulated by an industry serving the entire community.
Based on these invalid comparisons, the Commission projects a 50 per cent reduction in labour input requirements, taking 1991 as a starting point. No offsetting increase in other costs, such as payments to contractors, is allowed for. This would, if achieved, be a truly remarkable outcome. Labour productivity in the coal-dominated Australian industry is projected to surpass that of highly efficient US firms such as Consolidated Edison and Duke, which have a substantial nuclear component.
The Commission also assumes a 10 per cent reduction in capital construction costs. The evidence for this is derived from a secondary source, and is not very strong. Finally, the Commission treats cost savings arising from increased use of gas-fired stations as a benefit of reform. It claims that previous reliance on coal was the product of political decisions reflecting, among other things, a view that gas was too valuable a fuel to be used in power generation. On the basis of price expectations in the early 1980s, when the last big construction decisions were made, this was a perfectly reasonable view. The recent swing to gas reflects the fact that real interest rates, and hence capital costs, are much higher, and fuel costs much lower, than was anticipated in the early 1980s. In addition, the move to gas-fired electricity generation is apparently based on projections to the year 2020, which are subject to obvious uncertainty and are quite inconsistent with the timescale for the rest of the exercise.
The Commission's analysis of microeconomic reform in the water industry is broadly similar to its analysis of the electricity industry, but introduces one new error in benchmarking. The Commission adopts the invalid practice of using one authority as its best practice benchmark for labour productivity and another as the best practice benchmark for capital productivity. This must reflect different capital-labour ratios, since otherwise the same firm would be best on each measure. Hence, the resulting benchmark will lie outside the production possibility frontier.
In deriving alternative estimates of the benefits of microeconomic reform for the electricity, gas and water sector, it is assumed that the extra total factor productivity gain, estimated by the Economic Planning Advisory Council (1994) at a rate of 1 per cent per year, continues for four years, yielding a net gain of 4 per cent in productivity. This is equivalent to 0.1 per cent of GDP (Table 1, Column 2).
2.5 Rail, road and ports
The reforms in the transport sector considered by the Industry Commission (1995a) include moves to best practice in the railways and ports, and changes in road pricing policy. The Commission projects productivity gains of 15 per cent for railways, 14 per cent for ports and 5 per cent for road freight transport. These productivity gains imply a direct benefit equal to 0.18 per cent of GDP (Table 1, Column 1).
Unlike other areas where the starting point for reform is taken as early as 1990, the Commission attempts to use the current situation as the starting point for reform of rail and ports. This is surprising, since, unlike the areas discussed above, it appears that the removal of restrictive work practices and overstaffing has led to genuine productivity gains in the past.
The projected future gains in rail transport are based on the estimate that a move to world best practice will generate a 20 per cent reduction in costs. Although it does not appear that the projected reforms are sufficient to generate such a cost reduction (particularly as regards capital productivity), this estimate will be accepted as the basis for subsequent analysis, as will the estimated 14 per cent gain in port productivity. In the case of road transport, the only projected policy changes involve alterations in registration charges. An estimated 'notional' productivity gain of 5 per cent is arbitrarily assumed by the Commission without any basis (except for the fact that a study by Swan Consultants (1994) arbitrarily assumed a gain of 10 per cent). This gain, which would amount to about 0.1 per cent of GDP, is not included in subsequent projections. Hence, the alternative estimate is that microeconomic reform will generate direct benefits equal to 0.08 per cent of GDP (Table 1, Column 2).
2.6 Competitive tendering and contracting
The Industry Commission (1995a) examines the possibility that a large proportion of the services provided by governments could be subjected to competitive tendering and contracting (CTC). On the basis of an optimistic estimates of the scope for CTC and an assumption that an average cost saving of 20 per cent will be achieved, the Commission estimates direct savings of $2.7 billion or nearly 0.7 per cent of GDP from compulsory competitive tendering of current expenditure and $1.2 billion or 0.3 per cent of GDP from tendering of capital expenditure. However, in recognition of the scope for error in these estimates, only half of these gains are incorporated into the Commission's analysis, implying a direct benefit equal to 0.5 per cent of GDP (Table 1, Column 1).
The estimates presented by the Industry Commission (1995a) have subsequently been modified in the Draft Report on Competitive Tendering and Contracting (Industry Commission 1995b). The principal source of divergence between the estimates was the acceptance by the Industry Commission (1995b) of the point, made by Quiggin (1994) that some of the budgetary cost saving from CTC represented a transfer from employees in the form of reduced wages and increased work intensity. Although some important difficulties remain (Quiggin 1996), the analysis presented by the Industry Commission (1995b) represents a considerable advance on that of the Industry Commission (1995a).
The Industry Commission (1995b) reported the results of eight simulations of the effects of a program of CTC. The simulation presented by the Commission as Scenario 4B takes account of most of the points raised by Quiggin (1996), and will be used as the basis for the alternative estimate of the direct benefits of CTC. Under this scenario, the Industry Commission (1995b) projects that productivity gains from competitive tendering and contracting will be equal to 0.42 per cent of the total recurrent expenditure of government and government trading enterprise. This implies a direct benefit equal to 0.13 per cent of GDP (Table 1, Column 2).
2.7 Statutory marketing arrangements
The main reform considered for the agricultural sector is the removal of statutory marketing arrangements for dairy products, sugar and tobacco. The Industry Commission (1995a) estimates a final benefit equal to 0.15 per cent of GDP (Table 1, Column 3), or more than $600 million, from these reforms.
No separate estimate of direct benefits is presented in this case. However, the Commission's estimates of the final benefit may be compared to the results of a standard partial equilibrium analysis. For dairy products, the Commission estimates that the arrangements are equivalent to a 20 per cent tax, resulting in a transfer of $230 million from consumers to producers. Harberger's (1964) triangle estimate of the welfare costs of a tax is:
D = 0.5 e t2
where D is the cost of the tax, expressed as a proportion of total output, e is the sum of the elasticities of demand and supply and t is the tax rate. Assuming e = 2, this yields a welfare loss of $46 million from the dairy arrangements. Similar reasoning suggests a loss of $21 million from the tobacco arrangements. The case of sugar is more complex, as the regulations include restrictions on land use. It will be assumed that, under existing arrangements, the effective tax revenue of $24 million arising from higher sugar prices is entirely dissipated, with no gain to producers. Thus, the benefit of removal of sugar industry regulation is estimated at $24 million.
On the basis of this partial equilibrium analysis, the final benefit of microeconomic reform arising from the removal of statutory marketing arrangements for dairy products, tobacco and sugar is estimated at $91 million or 0.02 per cent of GDP (Table 1, Column 4).
2.8 The professions
The reforms to the medical and legal professions examined by the Industry Commission (1995a) involve the removal of various restrictions on competition. These are projected to generate productivity improvements of 50 per cent in the legal profession and a cost reduction of 0.3 per cent for the health sector as a whole. These estimates imply a direct gain equal to 0.10 per cent of GDP (Table 1 Column 1).
The Commission's estimates of productivity gains for the legal profession are an example of the misuse of upper bound estimates. The Commission cites evidence showing that the introduction of competition in conveyancing in the United Kingdom was followed by a reduction in conveyancing charges from 0.6 per cent of a property's value in 1984 to 0.4 per cent in 1986. Since the period in question was one of booming property values, the real reduction in charges was, at most, one-third.
In discussing the abolition of the functional separation between solicitors and barristers, the Commission cites the observation of Moran and Burns (1992) that, under functional separation, two lawyers are required where the services of one might be adequate, and concludes that cost reductions of 'up to 50 per cent' might be feasible. This upper bound calculation is the basis of an assumption that competition will generate 50 per cent cost reductions in the cost of all legal services provided by barristers and all conveyancing services. This corresponds to a reduction of 1.5 per cent in the cost of business services.
In deriving an alternative estimate of the benefits of reform in the legal profession it is assumed that competition will generate a 33 per cent reduction in the cost of conveyancing and barristers services, rather than the 50 per cent assumed by the Commission. The Commission's projections of efficiency gains arising from reform of the medical profession are accepted. These assumptions yield an alternative estimated direct benefit equal to 0.06 per cent of GDP (Table 1, Column 2).
2.9 Building regulations
The Industry Commission (1995a) suggests that the Hilmer recommendations on anti-competitive regulation should be extended to building regulations which it regards as unnecessarily stringent. As an example, it quotes claims made by the Department of Immigration, Local Government and Ethnic Affairs (1989) for the that reductions in the width of pavements and nature strips can be made 'without loss of safety or amenity'. Direct savings of $100-$150 million a year from this source and of $350 million a year in total (0.09 per cent of GDP) are estimated. A further saving of $750 million (0.18 per cent of GDP) is projected to arise from the elimination of unwarranted delays in the approval of building proposals, resulting in a total direct benefit equal to 0.27 per cent of GDP (Table 1, Column 1).
Cost savings arising from reductions in quality must be discounted and may be negative. No doubt there are benefits to be obtained from relaxing regulations in some areas, and from tightening them in others, and the normal processes of government should include continuous review of the regulatory setup. But the Commission makes no case for a general relaxation in regulations.
The Commission's observation that 'unwarranted delays' in gaining approval for building projects fail the public benefit test is true by definition. Equally, 'warranted delays', by definition, pass the public benefit test. The Commission provides no basis for distinguishing between warranted and unwarranted delays. It observes that one local council has achieved a median processing time of 28 days compared to the New South Wales median time of 49 days. However, the estimates used by the Commission are based on the assumption that all regulation is unnecessary. Making the still optimistic assumption that a general reduction of 40 per cent can be achieved (equal to that given in the Commission's single piece of evidence) the net saving is reduced to $300 million a year, which is equivalent to 0.08 per cent of GDP (Table 1, Column 2).
2.10 Private monopolies
The reforms to private monopolies examined by the Industry Commission (1995a) are the removal of restrictions on entry to the taxi industry and the prohibition of exclusive delivery arrangements negotiated between local newsagents and some newspapers. No direct estimate of benefits is available.
The monopoly associated with restrictions on entry to the taxi industry is estimated to represent an impost of $320 million a year on consumers. Most of this is capitalised into the price of taxi plates, so that the removal of restrictions would in large measure constitute a transfer from taxi plate owners to consumers rather than a net social welfare gain. The net gain would presumably be well below $320 million, but the simulations yield an estimated final benefit of $460 million or 0.14 per cent of GDP (Table 1, Column 3).
Assuming that the elasticity of demand for taxi services is 1 and that an average fare is $10, so that the entry restriction is equivalent to a 20 per cent tax, standard partial equilibrium techniques yield a net welfare gain of $60 million from the elimination of restrictions on entry to the taxi industry. Similar analysis applied to newspaper delivery arrangements suggest a gain of $75 million. Hence, the alternative estimate of the final benefit of reform of private monopolies is equal to $135 million or 0.03 per cent of GDP (Table 1, Column 4).
The Industry Commission (1995a) examines the replacement of external regulatory controls by self-regulation. The Commission assumes that about 10 per cent of the economy could be moved towards self-regulation, and that on average such a move would generate a productivity gain of 1 per cent in the sectors affected. These assumptions imply a direct benefit equal to 0.1 per cent of GDP (Table 1, Column 1).
The examples given are those of meat inspection and defence procurement. Failures of the self-regulation system in meat inspection are not discussed. The Commission observes gains of $4 million from the removal of various charges for meat inspection, and assumes, without any evidence, that the long run gains will be between 5 and 10 times this amount.
In the meat inspection case, the Commission takes account of the costs of the private inspectors who replaced public inspectors. In the case of defence procurement, no such allowance is made. A claimed savings of 600 positions is converted, without explanation, into a saving of $50 million. It would seem more reasonable, though still optimistic, to assume that these 600 public positions were replaced by 300 private positions, yielding a savings of 300 positions. Assuming salary and related costs of $40 000 per person, the net savings is $12 million.
The observed savings from self-regulation in the meat industry come to 0.15 per cent of the total output of the industry and the savings in defence procurement computed above amount to 0.25 per cent of the total costs of defence procurement. Assuming that productivity gains of 0.2 per cent could be achieved for that 10 per cent of the economy which, the Commission argues, could be subject to self-regulation, a net gain equal to 0.02 per cent of GDP would be generated (Table 1, Column 2). Against this must be set any losses associated with declining standards of quality under self-regulation. The experience of deregulation of the meat, airline and banking industries all suggest the reality of this possibility. Nevertheless, no deduction for declining quality standards is incorporated in the alternative estimates.
3 Sensitivity analysis
Any analysis of a program as broad as microeconomic reform must involve a large number of assumptions, and sensitivity analysis is therefore desirable. The Industry Commission (1995b) provides such an analysis in its consideration of the economic benefits of competitive tendering and contracting. However, the Industry Commission (1995a) does not report a sensitivity analysis, except insofar as the reporting of consumption growth measures allows a partial correction for the effects of the Commission's closure assumptions on the growth of capital stock and GDP.
Since the estimate of benefits of microeconomic reform derived in the present paper is below that reported in other studies, attention will be confined to an assessment of alternative assumptions that would result in higher estimates of total benefits. Three variations to the assumptions of the present paper are considered.
The first change is to the counterfactual assumption regarding productivity growth in the government business enterprise sector in the absence of microeconomic reform. Instead of supposing that previously observed productivity growth rates would remain unchanged, suppose that the annual rate of productivity growth would decline by one percentage point in the absence of reform. Assuming a five year timeframe, the absence of reform would therefore imply a cumulative productivity loss of 5 per cent for the government business enterprise sector, or around 0.25 per cent of GDP. Taking account of unemployment effects, the final benefit estimate would be increased by around 0.22 per cent of GDP.
The second change is to use the Industry Commission Scenario 4A estimate of the direct productivity benefits of competitive tendering and contracting, which is based on a 20 per cent average reduction in budgetary costs, in place of the more conservative Scenario 4B, which incorporates a 10 per cent reduction. This would increase the estimated direct benefit of microeconomic reform from 0.13 per cent of GDP to 0.26 per cent of GDP, and the estimated final benefit from 0.11 per cent of GDP to 0.22
Third, consider the exclusion of effects on unemployment. In a fully employed economy with no externalities, the net social gain from a productivity increase in a given industry will be equal, on average, to the change in producer and consumer surplus in the industry concerned, and may be estimated, up to a second-order approximation, by the product of the proportional increase in productivity and the value added in the industry. Hence, if unemployment is disregarded, the benefit of productivity-increasing microeconomic reform will be equal to the direct benefit reported in Table 1, Column 2. The benefits from removal of monopoly pricing distortions and reform of statutory marketing will be equal to the final benefit reported in Table 1, Column 4.
Considered in isolation, the exclusion of unemployment effects would increase the estimated benefits of microeconomic reform to 0.74 per cent of GDP. In combination with the alternative assumptions on counterfactuals and the benefits of competitive tendering and contracting, this change would yield a final estimate of benefits equal to 1.12 per cent of GDP.
The combined effect of this more optimistic set of assumptions is to nearly double the estimated benefit presented in Table 1. The resulting estimate of direct benefits is approximately half the Industry Commission's (1995a) estimate of direct benefits equal to 2.32 per cent of GDP. The range of disagreement could be reduced further if the Commission's (1995a) estimates of the benefits of competitive tendering and contracting were replaced by the more conservative (1995b) estimates.
The final benefit estimate is still well below the Industry Commission of 5.46 per cent, reflecting the fact that more than half of this estimate arises from the choice of closure assumptions. The arguments presented above show that the GDP increase arising from the closure assumptions would have no welfare relevance even if those assumptions were satisfied.
4 Other studies
A number of other studies (Industry Commission 1990; Bureau of Industry Economics 1990; Business Council of Australia 1994; Dao and Jowett 1994; Filmer and Dao 1994) have examined the benefits of microeconomic reform. All have taken some version of the ORANI model as a starting point, and all have examined similar sets of reforms and assumed roughly the same direct productivity impacts. In the absence of further adjustments, all estimate the benefits of microeconomic reform to be around 6 per cent of GDP, a value similar to that obtained by the Industry Commission (1995a).
The key difference between the studies is that, in addition to the optimistic assumptions about the direct impact of reform made by the Industry Commission, the authors of the other studies assume, without supporting evidence, that microeconomic reform will generate increases in output arising from across-the-board productivity increases or reductions in the 'natural' rate of unemployment.
For example, the Bureau of Industry Economics (1990) assumed a doubling of the rate of productivity growth. Extrapolated over ten years, this yields a GDP gain of 16 per cent. The Business Council of Australia (1994) and Filmer and Dao (1994) made more modest assumptions about productivity gains, ranging from 3.5 to 7 per cent, but compensated for this by assuming an increase of around 5 per cent in total employment. The assumed employment gains are particularly striking in view of the fact that the direct effect of most of the reforms modelled is to reduce employment. With these adjustments, the projected benefits of microeconomic reform range from 12.7 per cent of GDP (Filmer and Dao 1994 ) to 21 per cent (Bureau of Industry Economics 1990 and Business Council of Australia 1994).
All the studies discussed above are based on the Industry Commission version of the ORANI model, with similar assumptions They are not independent analyses, but variations on a common theme. Most of the criticisms set out above in relation to the Industry Commission (1995a) apply with equal force to each of them.
A check on the plausibility of projections of the benefits of microeconomic reform can be obtained by considering previous and projected future changes in macroeconomic variables. The growth rate of the labour force is around 2 per cent a year. If we assume that the trend increase in capital stock per worker and exogenous technological change will continue to yield an annual rate of labour productivity improvement of 1.5 per cent in the absence of microeconomic reform, the sustainable long term GDP growth rate is 3.5 per cent. Higher rates can be achieved only when unemployment is falling.
Using these parameters we may first consider whether microeconomic reform since 1986 has generated significant benefits thus far. The average GDP growth rate over the past ten years has been around 3 per cent. The shortfall from the long term growth rate reflects rising unemployment associated with the 1989-91 recession. If unemployment levels had remained unchanged over the decade, the long term GDP growth rate of 3.5 per cent would have been realised, but not exceeded.
On the basis of this evidence, microeconomic reform has had no significant impact on economic growth so far. This conclusion is accepted by advocates of microeconomic reform such as Fane (1995). Past microeconomic reforms may have yielded small benefits, on the scale of those estimated in this paper. Such small benefits would have no discernible impact on the ten-year growth rate and would have been swamped by the 1989-91 recession. However, if past microeconomic reforms had yielded benefits of the order of 5 per cent of GDP, as estimated by the Industry Commission (1990), this should have been evident in the growth performance of the economy.
Hence, any substantial benefits from microeconomic reform must lie in the future. It therefore seems reasonable to treat the results of all the studies mentioned above as projections of the additional GDP growth that would be generated by microeconomic reform over the five-year period commencing in 1995. Thus, the Industry Commission's (1995a) estimates would call for additional growth of 1 per cent, implying a sustainable growth rate of 4.5 per cent. The estimates of Filmer and Dao (1994) and Dao and Jowett (1994), which assume a reduction in unemployment, require annual average growth rates of 6 per cent. Finally, the Business Council of Australia's (1994) estimates require growth rates of 7.5 per cent. By contrast, the estimates offered in this paper suggest that microeconomic reform will continue to have no discernible impact on the rate of economic growth.
Current macroeconomic projections such as those in the 1996-97 Budget Papers reflect an implicit assumption that microeconomic reform will have little or no impact on economic growth, productivity, or on the level of unemployment. These projections are based on the assumption that the sustainable rate of GDP growth is around 3.5 per cent. The natural rate of unemployment is implicitly assumed to be between 7 per cent and 8 per cent after taking into account the anticipated effects of labour market reform. These macroeconomic projections are consistent with the estimates of the effects of microeconomic reform derived in this paper, but not with the estimates presented by the Industry Commission (1995a) and related studies.
5 Concluding Comments
Australians have been promised for at least a decade that the process of micro-economic reform will yield substantial improvements in living standards. No such improvements have been evident so far. Careful analysis of the Industry Commission (1995a) indicates that even a comprehensive program of microeconomic reform is unlikely to yield welfare gains that would be detectable over the noise generated by macroeconomic fluctuations in the economy. The direct gains from the entire program of microeconomic reform are likely to be no more 1 per cent of GDP and these may be partially offset by resulting increases in unemployment.
This conclusion is supported by the experience of countries such as United Kingdom and New Zealand, where reform has been more comprehensive than in Australia. Despite some periods of strong growth, neither country has outperformed the OECD average growth rate since reforms commenced in 1979 and 1984 respectively. Evans, Grimes and Wilkinson (1996), uncritical supporters of the New Zealand reforms, report that the average rate of per capita GDP growth for the period 1984-95 was 0.97 per cent, considerably below the rates of between 1.5 and 2 per cent achieved in Australia. After two years of good performance in 1994 and 1995, New Zealand's annual GDP growth rate has fallen to 2 per cent, consistent with maintenance of a trend rate of growth of per capita GDP around 1 per cent. The United Kingdom has performed somewhat better but not sufficiently well to catch up to other European countries that have not undertaken reform, or even to Australia.
None of this implies that microeconomic reform should be rejected outright. Many individual reforms may have net social benefits. However, the notion that microeconomic reform is a process that must be pursued regardless of the short term costs should be recognised as a mistake. Reforms should be assessed on a case-by-case basis, with careful attention being paid to the consequences for consumers, employees and society as a whole.
Aigner, D., Lovell, C. A. K. and Schmidt, P. (1977), 'Formulation and Estimation of Stochastic Frontier Production Function Models', Journal of Econometrics 6(1), 21-37.
Arena, A., Bahtesevanoglou, J., Branton, S., (1992), The importance of telecommunications in Australia's economic development and the likely effects of telecommunications reform, Austel, Occasional Paper, Economics 1.
Bureau of Industry Economics, (1994), International performance indicators -- electricity, Research Report 54, AGPS, Canberra.
Bureau of Industry Economics, (1990), Microeconomic Reform and the Structure of Industry, BIE Discussion Paper No 9, AGPS, Canberra.
Business Council of Australia, (1994), Business Council Bulletin, , Melbourne.
Butlin, N., Barnard, A. and Pincus, J. (1982), Government and Capitalism: Public and Private Choice in Twentieth Century Australia, Allen and Unwin, Sydney.
Cubbin, J., Domberger, S. and Meadowcroft, S. (1987), 'Competitive tendering and refuse collection: Identifying the sources of efficiency gains', Fiscal Studies 8(3), 49-58.
Dao, D., Jowett, T., (1994), Economic effects of microeconomic reform -- supplementary paper, , EPAC Working Paper.
Department of Immigration, Local Government and Ethnic Affairs, (1989), Local government regulation of land and building development , , Study undertaken by the University of Tasmania.
Economic Planning Advisory Commission Private Infrastructure Task Force, (1995), Interim Report, AGPS, Canberra.
Economic Planning and Advisory Council (EPAC), (1994), Microeconomic reform paying off for Government Business Enterprises, AGPS, Canberra.
Evans, L., Grimes, A. and Wilkinson, B., (1996), Economic reform in New Zealand 1984-95: The pursuit of efficiency, Journal of Economic Literature 34(4), 1856-1902.
Fane, G., (1995), Economic reform and deregulation in Australia, Discussion Paper No. 319, Centre for Economic Policy Research, ANU.
Filmer, R., Dao, D., (1994), Economic effects of microeconomic reform, AGPS, EPAC Background Paper no 38.
Forsyth, P. (1992), 'A perspective on microeconomic reform', in Forsyth, P. (ed.), Microeconomic Reform in Australia, Allen and Unwin, St. Leonards, NSW.
Harberger, A. (1964), 'Taxation, resource allocation and welfare', in NBER (ed.), The role of Direct and Indirect Taxes in the Federal Revenue System, Princeton University Press, Princeton.
Hilmer, F., Rayner, M., Taperell, G., (1993), National Competition Policy, Report by the Independent Committee of Inquiry
Industry Commission, (1990), Annual Report, 1989-90, AGPS, Canberra.
Industry Commission, (1995a), The growth and revenue implications of Hilmer and related reforms, AGPS, Canberra.
Industry Commission, (1995b), Draft Report on Competitive Tendering and Contracting by Public Sector Agencies, AGPS, Melbourne.
Moran, A. J., and Burns, G., (1992), Monopolistic restrictions in the provision of advocacy services, Tasman Economic Research.
Quiggin, J. (1996), 'Competitive tendering and contracting in Australia', Australian Journal of Public Administration forthcoming.
Quiggin, J. (1994), 'The fiscal gains from contracting out: Transfers or efficiency improvements?', Australian Economic Review September, 97-102.
Schmidt, P. (1976), 'On the Statistical Estimation of Parametric Frontier Production Functions', Review of Economics and Statistics 58(2), 238-39.
Swan Consultants, (1994), Impact of road transport diesel taxes on the Australian economy, , Report prepared for the Road Transport Forum, Canberra.
Go back to John Quiggin's home page
Go back to John Quiggin's 1997 articles