Positive and Normative
It is a matter of fact that governments attempt to influence the actions of industrial companies with a range of objectives in mind. Some illustrations make this clear. Governments frequently express concern about the rate of growth of the economy and seek to influence the rate of growth of industrial output by measures designed to raise the level of industrial investment and to improve the efficiency of factor inputs, for example by supporting technical education and research and development expenditure. They respond to political pressures from different regions by measures designed to iron out differences in economic growth between them, for example by allocation of government purchases. The present-day concentration of interest on the ‘diseconomies’ created by environmental pollution has produced a response by government in the form of anti-pollution regulation and the use of fiscal measures.
A positive approach to the study of government industrial policy seeks to identify the policy objectives of government which are actually operative and to use economic analysis to examine whether or not the measures adopted achieve the stated objectives. No judgement is passed on the objectives themselves. A normative approach evaluates the objectives themselves. A set of policies are classified as ‘good’ or ‘bad’, according to whether or not they conform to a particular ‘norm’. Thus economists frequently argue that satisfying consumer wants is the main function of the economy and that industrial policy measures ought to conform to and therefore are to be judged by that aim. However, while disagreements over the effects of a government measure can be resolved by appeal to facts and logic, in principle at least, whether the measure is ‘good’ or ‘bad’ is a matter of personal opinion. The fact that there may be a consensus view amongst economists identifying some policies as ‘good’ does not mean that ‘good’ policies can be derived from economic expertise.
In practice, it has to be accepted that a separation between normative and positive approaches to industrial policy may be difficult. It is often not possible to prove beyond question that a particular policy measure will achieve its objective; in which case analysts may be biased, consciously or unconsciously, in favour of the evidence which supports their normative position. Again, one's normative position might be influenced by positive analysis. A strongly held view that a high rate of economic growth is desirable, might be modified if it were shown that it entailed the use of industrial policy measures which made the distribution of income more unequal.
Determinants of Government Expenditure Growth
In answering this question, it is assumed that all the variables can be identified and clearly defined. The dependent variable, government spending, could be defined in a number of ways. Here it is taken to refer to expenditure on goods and services and transfers and refers to all layers of government: public enterprises are excluded. This definition influences the choice of the independent variable. For example, the proportion of government employees in total employment, which is taken to include employees in all layers of government, but excludes those employed in public enterprises. The remaining independent variables may be assumed to be capable of clear definition.
A further preliminary point is to establish the relevant time period, for the influence of the independent variables may be different in the long and in the short run. It is assumed in this answer that the time period is determined by the availability of reliable statistical data which suggests concentrating on the last thirty years.
The commonsense view would be that the growth in population would increase the demand for government services: more children means more schools, larger working population means more transport facilities and social benefits, more retired people means more pensions. This view neglects the possibility that there could be a ‘scale effect’ – the extra output of government services need not require a proportional increase in inputs. It would also have to be modified in the light of the speed of population growth. For example, slow-growing populations are associated with an ageing of the population. The connexion between (p) and (G) then becomes more complicated. The growth in the demand for educational resources could fall, whereas the growth in the demand for pensions and social services for the aged would rise.
Historically, the income elasticity of demand for government spending with respect to GDP is positive, though this is not to say that government spending has risen purely in response to the rise in average incomes. Nevertheless, it would be reasonable to expect that a rise in living standards would be associated with a rise in standards of provision of government services including increases in the real income of transfer receivers.
A growing proportion of working population employed by government might have two positive effects on government spending. The first is that government employees presumably have a vested interest in government spending and their growth in numbers relative to the size of working population makes them a growing proportion of the electorate. The second is that they are under less pressure to be ‘cost conscious’ if they are producing services which are not priced. Empirical evidence is sparse.
Natural Monopoly
An industry is said to be a natural monopoly if a single producer can supply the entire market at lower total cost than any conceivable combination of two or more producers whose aggregate output sums to the same quantity as the single supplier. An industry for which this condition holds is said to have a cost structure which is ‘sub-additive’. More formally this can be expressed as follows:
for all qi, qj, qk, such that Q=qi+qj+…+qk, where C(qi) represents the total cost of producing output qi, etc. The most obvious situation in which this condition will be satisfied is when average costs of production decline continuously because of economies of scale. However, such continuously declining average costs are not necessary for a natural monopoly to exist even though in the single product case they are sufficient.
Natural monopoly is most often associated with public utilities such as gas, electricity and water, or with the area of public transport such as the railways. In these industries large fixed costs are incurred in laying down networks of pipes, cables, wires, or permanent ways. Costs are therefore expected to decline over fairly large ranges of output.
Contestable Markets
A market is perfectly contestable if it is impossible for incumbent firms to leave a profit opportunity without a new entrant taking advantage of it. Contestability represents a situation of ‘ultra-free’ entry into, and exit from, an industry. Any attempt by a monopolist to make high profits will be instantly thwarted in a perfectly contestable market by an entrant supplying the whole or part of the market at a lower price. The incumbent simply loses customers who transfer to the new supplier.
The condition of costless exit from an industry implies that there can be no sunk costs incurred upon entry. All costs can be recovered. Physical capital can be costlessly transferred to some other use, or sold on the market for its original value less any economic depreciation. In these conditions an entrant always faces the same cost calculations as an incumbent, even if the length of time the entrant plans to serve the market is extremely short. The industry is vulnerable to ‘hit and run’ entry.
Entry will not occur, even in the conditions described above, if a potential entrant expects the incumbent to respond instantly by adjusting his price to match that of the new entrant. For a market to be perfectly contestable therefore, it is necessary that there exists some positive time-lag in the response of the incumbent, during which the new entrant can reap the benefits of entry into the market.
Although the conditions for perfect contestability will not in practice be satisfied in any industry, the force of potential competition in disciplining incumbent firms will often be important. Conditions in the civil air transport industry have been considered to be contestable by some analysts.
RPI − X
RPI − X refers to a regime of price regulation devised for natural monopoly industries. Under this system, the price of the product cannot rise by more than the rate of increase of the retail price index (RPI) minus some factor ‘X’ to take account of expected productivity improvements in the regulated industry. Clearly ‘X’ is subject to negotiation, and will change from time to time. If adverse changes (such as the rise in price of some important input) are possible, they may be taken into account by adjusting the formula to RPI + Y − X, where Y is the rate of increase in the price of the important input.
This regime of regulation was first introduced in the UK in the case of British Telecom, and in the form of RPI + Y − X in the case of British Gas. The advantage over regimes of regulation based on profits, rates of return or costs, is that the firm has an incentive to pursue productivity improvements knowing that these will not be followed immediately by more stringent price regulation. However, in practice, the factor X must be reassessed at intervals, so that in the long term a high rate of productivity improvement might be likely to result in a higher value of ‘X’ and hence a more stringent regime.
The Demsetz Auction
The Demsetz or Chadwick auction is a method of introducing competition into areas of natural monopoly by using competition for the market to make up for the absence of competition in the market. Potential suppliers of a good or service would bid for the franchise. They would do so not by offering a lump sum or a royalty to the government in return for the right to act as a natural monopoly supplier, but would instead bid by quoting prices at which they were prepared to supply the market. Other things equal, the supplier quoting the lowest price would be offered the franchise. In principle, competing bidders would drive the price down to a competitive level and no monopoly profit would be made by the eventual producer.
Theoretically attractive, the Demsetz auction faces considerable problems of implementation including defining product quality in the franchise contract, coping with failure to comply with the contract conditions, setting a suitable length of contract, valuing capital at contract termination, and adjusting to changes in conditions over time. Nevertheless the mechanism is being used in the area of some local authority services, and in varying degrees of purity in television broadcasting and cable television.
First calculate the net present value (NPV) of the two projects in the absence of regulation
| Project I | |
| Labour costs per year | £m. |
| (employment x average wage) | 2.5 |
| Maintenance and raw materials costs per year | 1.5 |
| Hence: operating costs per year (Cl) | 4.0 |
| Revenue from sales per year (Rl) | 5.0 |
| Capital investment (Kl) (incurred immediately) | 5.0 |
| Thus (over an infinite time horizon) | |
| NPV(1) = −K1+(R1−C1)/0.1=5, where the rate of discount is 0.1. | |
| Project H | |
| Labour costs per year | £m. |
| (employment x average wage) | 1.25 |
| Maintenance and raw materials costs per year | 2.35 |
| Hence: operating costs per year (C2) | 3.60 |
| Revenue from sales per year (R2) | 5.00 |
| Capital investment (K2) (incurred immediately) | 9.33 |
| Thus (over an infinite time horizon) | |
| NPV(2) = −K2+(R2−C2)/0.1 = 4.67, where the rate of discount is again 0.1. | |
Clearly, in the absence of regulation, Project I is preferred to Project II. NPV(1)−NPV(2) = £0.33m.
Now consider the impact of regulation:
Project I
In the case of Project I, the observed rate of return on invested capital will be (R1 − C1)/K1 = 0.2. This 20 per cent return exceeds the permitted maximum, and the regulators will reduce the price of output to £4.75 per unit in order to restrict the company to a 15 per cent return on capital invested. R1 will thereby be reduced to £4.75m. and the NPV of the project in the presence of regulation NPV’(1) will be £2.5m., i.e. NPV‘(1)=−5+(4.75−4.00)/0.l=2.5
Project II
In the case of Project II, the observed rate of return is 15 per cent even without the interference of regulators. The excess of revenue over operating costs in each period is £1.4m. and this is precisely 15 per cent of the £9.33m. invested. Thus the regulated price of the output would remain unchanged at £5. It follows that the NPV of the project to the firm is unchanged by regulation. NPV‘(2) is £4.67m.
Conclusion
The regulatory system will indeed affect the investment decision of the firm. The firm will prefer Project II under regulation but Project I when not constrained by the regulators. This result is consistent with the Averch-Johnson effect. Project 11 involves a greater investment of capital, greater maintenance expenditures, but less reliance on labour compared with Project I. Both projects result in the same output, but their cost structures differ. The more capital intensive cost structure is preferred under rate of return regulation.
Asymmetric Information
In examining economic relations between government and a firm, it has become common practice to regard the government as the ‘principal’ which enters into an agreement with the firm, as its ‘agent’, to carry out some agreed courses of action. In principle, a contract can be drawn up between the principal and agent, but, in practice, it may not be possible for the government to detect if the contract is reasonably specified or, is being adhered to once specified, if the agent's activities are not perfectly observable. This is called the ‘asymmetric information problem’, for the problem is how far firms are tempted to obtain more generous terms in the contract than would be possible if its activities could be closely monitored, and how far the government, as principal, will wish to invest resources in improving information in order to enforce the contract more completely.
The problem of asymmetric information can be exemplified in a number of different situations, (four) of which are now considered.
Government purchasing contracts
The existence of the problem depends on the market environment. If the government is a monopsonist faced with competing suppliers of a clearly specified product, competition between firms will ensure that they will have an incentive to reveal information about their activities in order to obtain the contract. Where the government is faced with only one seller, the government cannot compare the offers made by one seller against another and has to rely on the sole seller for information on its costs which will form the basis of the sales contract. The government purchaser may be faced with inflated cost estimates and checking these requires the use of resources which have alternative uses, e.g. post-costing of contracts and delay in final payments to the seller until final cost estimates are agreed.
Selective government subsidies to industry
Such subsidies are usually given to alter the value of some ‘target’ variable of interest to the government such as an increase in a firm's investment, output or employment. The precise link between the amount of subsidy in the movement of the target variable(s) in the desired direction depend on information supplied by the firm. For example, the firm's incentive to increase output will depend on whether its profits will increase, but profit estimates will be a function of the information supplied to the government about the change in costs incurred in increasing its output to the desired level. The firm is in a position to exercise strategic behaviour not only because it is the sole source of information about costs but also because the firm itself may not have a precise idea of the costs that it will incur.
Government bureaucracy
A government ‘bureau’ may be regarded as a ‘firm’ supplying services to the government in power though it does not price its output. A government allocating a budget for a particular service may be assumed to wish for ‘value for money’, in the sense of obtaining the maximum output from a given budget. This would entail minimizing the cost per unit of output. However, the sole source of information on the nature of output and the ‘production function’ is the bureau and, as bureaux are prevented from making profits, the incentives to minimize costs per unit of output are weak. The principal is again faced with the problem of whether or not to invest resources in checking on the information on cost per unit of output provided by the agent.
Control of environmental damage
Regulation of environmental damage caused by pollution, for example, can take several forms, but the asymmetric information problem is clearly demonstrated in the case where government agencies rely on enforcement of regulations. Successful regulation depends on the ability to calculate the damage from a particular pollutant which entails estimating both the physical effects and their translation into ‘welfare loss’. The physical effects of the pollution of any one plant may be uncertain and can depend on climate and position and the existence of other pollutants. The costs of corrective measures undertaken by firms may be uncertain. Experts may not be able to agree amongst themselves on the measurement of both effects. Firms will have both the incentive and the opportunity to bargain with government regulators about the enforcement of anti-pollution measures and can rely on the asymmetric information problem to raise the costs to the regulators of obtaining agreement about the anti-pollution measures, the costs of which have to be incurred by the firms themselves.
Monopolistic and Bureaucratic Behaviour
The standard analysis of the economics of monopolistic behaviour and that of bureaucratic behaviour draws a contrast between the motivation of the monopolist and bureaucrat respectively. The former maximizes profits but the latter is precluded by the contract of service from appropriating any surplus over and above the cost of producing any given level of output. Accordingly, the motivation of the latter is usually taken to be that of maximizing the size of the budget of the bureau. These motivation assumptions are rather extreme, but serve as a first approximation in exploring the behaviouristic parallel.
The diagram shows that both monopolistic and bureaucratic behaviour lead to a deviation from the ‘social optimum’, which is taken to be the equation of marginal cost with the ‘price’ or ‘marginal evaluation’ (Pc with output q2). However, using this criterion, the monopolist ‘under-produces’ by equating marginal cost with marginal revenue (with output q1<q2) whereas the bureaucrat ‘over-produces’ (with output qb where q2<qb ≤ q3). The limitation facing the latter is the point (q3) where the marginal evaluation of his output is zero. The bureaucrat exploits the existence of asymmetric information (see answer 3(a)) in order to obtain the maximum size of grant which is compatible with the credibility of the estimates submitted to those providing him with finance.
Figure A2.1 Bureaucracy and monopoly again
The contrast between monopolistic and bureaucratic behaviour in this analysis rests heavily on the behaviouristic assumptions. These are often questioned. The theory of the firm employed above does not distinguish ownership from control and therefore does not distinguish between shareholders' and managers' objectives. Managers exploiting monopoly power may seek to appropriate any gains by increasing their own emoluments, subject to the constraint of paying dividends which satisfy shareholders. Bureaucrats may eschew the prestige and power, and possibly promotion, associated with maximizing their output and use their control over information to enable them to inflate costs which makes their jobs more comfortable. In the technical jargon, managers and monopolies and bureaucrats may share the characteristic of being ‘X’ or technically inefficient, as well as promoting an allocation of resources out of keeping with the objective of maximizing consumer surplus.
In the long run even the simple presumption that monopolists' and bureaucrats' actions are out of line with consumers' wishes may offer an incentive to governments to devise methods for controlling their actions. How effective these are likely to be in modifying the behaviour of both parties is a complex matter beyond the scope of this analysis.
The Public Enterprise and the Bureau
A ‘pure’ public enterprise may be defined as an enterprise which produces some marketed output using assets that are vested in the state, or an agency of the state. The assets of a pure public enterprise are therefore collectively held, and there are no privately exchangeable rights to the profits deriving from the enterprise's activities. A bureau uses assets which are similarly vested in the state, but it does not market its output. The revenue of the bureau is received out of taxation and depends on the decisions of government ministers responsible to Parliament.
In practice, the distinction between a bureau and a public enterprise can become blurred. This is because the proportion of revenue generated from sales of the output of an enterprise will not necessarily be either zero or 100 per cent. A public enterprise which receives a significant flow of subsidy from the government for specific activities, or to cover general losses, becomes more bureau-like. Conversely, a bureau which begins to market specialized services becomes more like a public enterprise. The sale of weather forecasting services by the meteorological office, for example, could be seen as a transformation from a bureau type of economic entity to a public enterprise type of economic entity.
The Public Enterprise and the Private Enterprise
The main characteristic which defines a ‘private’ enterprise is the existence of exchangeable claims to the profits of the enterprise which may be held by private individuals. This definition implies that co-operatives and non-profit firms would not be classed as private enterprises. For example, a worker's cooperative might give its members rights to a share in the profits. These rights would not normally be exchangeable, however, and therefore a co-operative would not be considered a private enterprise. Similarly, a non-profit or charitable enterprise would not be considered part of the private enterprise sector, under the definition given above, because there are no exchangeable residual claims.
Once again, the distinction between public and private enterprise is not always perfectly clear cut. The state may hold a proportion of the shares of a public company thereby creating a ‘mixed enterprise’. At what precise point a private enterprise is transformed into a public enterprise by the state's acquisition of shares cannot be determined with perfect objectivity.
Comparing Productivity of Public and Private Enterprises
Productivity comparisons aim to show whether property rights matter. Is it the case that the absence of exchangeable property rights in a public enterprise depresses incentives and results in a lower level of productivity compared with a private enterprise? The main problems which arise in attempting to answer this question are as follows:
Somehow it is necessary to try to isolate the influence of differing property rights from all the other factors which may affect the performance of an enterprise. Thus it is necessary, in principle, to compare enterprises of the same size, in the same industry, using the same technology, facing the same degree of competition in the product market, in an attempt to see whether other things equal a different property rights structure results in a different level of performance. This is done on the, probably false, premise that factors such as size, technology and competition, are themselves not affected by the property rights structures characteristic of the enterprises which make up the industry. The study by Davies (1971) of two Australian airlines, and the study by Caves and Christensen (1980) of two Canadian railroads, are examples of attempts to compare public and private enterprise in situations which are as closely comparable as possible.
The second problem is that productivity comparisons require agreed measures of inputs and outputs. The output measurement problem can be particularly acute if the very purpose of public ownership is to change the quality or type of output from what it might otherwise have been under private enterprise. Further, it is always open to an enterprise to claim that unfavourable productivity measures simply reflect some underestimated or unmeasured component of their output. A private enterprise running a reliable luxury coach service may seem to have a lower productivity than a public enterprise running an unreliable and extremely basic one if the output is measured simply in terms of passenger miles travelled. Conversely a private supplier who pollutes the environment may appear more productive than a public supplier who invests in cleaner technology if the ‘output’ is not properly identified.
The Control of Public Enterprise
Principal-agent analysis can be applied to any situation in which one party (the agent) agrees to act on behalf of another party (the principal). If the interests of the agent and the principal diverge, and if the principal is unable at low cost to determine what effort the agent is supplying, the problem of devising a suitable incentive scheme arises. In the context of the public enterprise, managers of the enterprise can be viewed as agents, and the responsible minister in the government can be regarded as the principal.
An agreement between principal and agent will specify a fee schedule such that the reward to the agent will vary with the final ‘outcome’ achieved. The more the agent's reward depends upon the final outcome, the greater is the incentive to the agent to exert effort on the principal's behalf. In the case of public enterprises the problem is in defining a simple ‘outcome’ measure. In public companies, managers can be paid profit bonuses, given stock-options, and other benefits linked to profitability. Managers may also fear take-over by other more profitable firms, and they face the possibility of bankruptcy if profits fall too far. These incentive devices depend upon a simple measure of ‘success’ – the profits of the enterprise – and the existence of tradeable claims to the profits.
In the case of a public enterprise the specification of the desired outcome is more difficult. In cases where the enterprise has monopoly power, the maximization of profit is not likely to be acceptable to voters as a suitable objective. Linking managerial rewards to the achievement of gains in ‘economic efficiency’ might be attractive to voters and politicians in principle, but measures of such gains are extremely difficult to devise. The absence of a clear summary measure of ‘success’ led in the 1967 White Paper Nationalised Industries to the imposition of ‘rules’ on the operation of public enterprises. These rules included the marginal cost pricing rule, and an investment rule based upon the calculation of the net present value of a project at a given ‘social’ rate of discount. These rules could be defended as a means to achieving economic efficiency. However, there existed no way of knowing, given high monitoring and information costs, whether or not the management of an enterprise was complying with these rules. There was no clear incentive for managers to do so, and as it was possible to reconcile virtually any results with some plausible interpretation of the rules, they became discredited as a means of public enterprise control.
More recently the pricing and investment rules have been replaced by financial and productivity targets. These are observable and measurable but may direct managerial effort in ways which do not always correspond to the principal's interests. For example, financial targets can be achieved by exploiting monopoly power, and productivity targets can be achieved by ignoring dimensions of output which are not specified by the target.
Privatization
The privatization of assets occurs when collective property rights held by the state are replaced by exchangeable private rights. Examples include the flotation of British Telecom and British Gas.
Deregulation
Deregulation occurs when the direct intervention in economic decisions by government or its agents using non-fiscal instruments is reduced. The removal of control on the wellhead price of natural gas in the US is an example of deregulation.
Liberalization
Liberalization occurs when competitive forces are permitted to operate where they have hitherto been inhibited. Usually this will involve reducing restrictions on the ability of competitors to enter an industry, trade, or profession. The competitive determination of brokerage fees and the abolition of the minimum brokerage commission in the UK in 1986 is an example of liberalization.
Liberalization may often accompany deregulation (for example the abolition of restrictions on entry into long-distance coach services), but some measures of liberalization may require regulatory enforcement (for example facilitating entry into gas or electricity supply by using the network as a ‘common carrier’). Similarly, privatization may increase (as in gas and water) or decrease (as in British Steel) the importance of regulation.
Technical Efficiency
An enterprise is technically efficient if it is producing the maximum physical output possible with the available resources. It is a concept which requires knowledge of purely technological possibilities.
Productive Efficiency
An enterprise is productively efficient if it is producing a given output at the lowest possible total cost given the prevailing prices of inputs. Technical efficiency is a necessary but not a sufficient condition for productive efficiency.
Economic Efficiency
Economic efficiency requires that the price of the product in the market reflects its social cost of production at the margin. Social costs include all the costs incurred in the production process, including those experienced by ‘outsiders’ through pollution or other ‘spillover effects’.
Productive efficiency is a necessary but not a sufficient condition for economic efficiency. A monopolist, for example, may be productively efficient, but the price and output level chosen by the monopolist is unlikely to be economically efficient.
Company I uses 60 units of labour and 15 units of capital to produce 10 units of output. Using process C it should be possible when technically efficient to produce 10 units of output with 40 units of labour and 10 units of capital. The company is therefore not technically efficient. Total costs of £750 exceed the technically efficient level of £500.
Company II uses 20 units of labour and 20 units of capital to produce 10 units of output. This is the minimum input level required to produce 10 units of output using process B. The company is therefore technically efficient.
Total costs in Company II are £400. This is £100 less than could be achieved using process C, even if process C were operated in a technically efficient way. Thus, of the £350 cost advantage achieved by Company II over Company I, it is possible to attribute £250 to greater technical efficiency, and £100 to greater productive efficiency.
With labour costs of £300 and capital costs of £l50 Company I is still using 60 units of labour and 15 units of capital to produce 10 units of output. It is therefore as technically inefficient as ever, despite a drop in measured costs of production from £750 to £450. The £300 difference represents a redistribution of income away from labour rather than a gain in efficiency.
The new lower price of labour may, however, lead us to reassess our conclusion that Company II was initially more productively efficient than Company I. At the new factor prices, process B has no advantage over process C with respect to productive efficiency. If the new prices are assumed to represent the true costs incurred by owners of factor inputs, it is possible to argue that the lower costs initially associated with process B were illusory, and that no real efficiency gains accrued to the use of process B over process C.
