Until recently the debate surrounding the reorganisation and privatisation of most of the major UK utilities has not featured the quality of the service provided strongly. Consumer groups, such as the Consumers' Association, National Consumer Council, and the Electricity and Gas Consumer Committees have complained about service standards and rises in some indicators, recently gas complaints and the proportion of trains running late, have excited comment. However, most focus has been on the levels of prices and profits and quality of service has been prominent mainly when it has affected them, such as through creating a need to fund the investment required to raise water quality to prescribed European levels.

This is not entirely surprising. There is a strong commodity element in most utility products with comparatively little opportunity for variation in quality. There was a general suspicion that, while in the public sector, utilities had been run mainly by engineers subject to inadequate cost scrutiny who had "gold-plated" their industries and provided excessive quality. Therefore, with the exception of the railways (and perhaps telephones) where it was felt that public borrowing constraints had inhibited investment for quality, the emphasis was on price to the customer and cost to the exchequer rather than on the precise service provided.

However, there is a literature on quality in regulated monopolies, largely reliant on Spence (1975), which establishes a link with the form of price regulation. UK regulators, with the help of the Monopolies and Mergers Commission, have established a broadly agreed methodology for price control and are now paying more attention to quality.

This paper discusses the relationship of quality to price control methods, reviews some of the techniques presently being used in the UK and suggests a way forward.

The essential problem is that the value of an increase in quality to a monopolist is the increase in price which can be obtained times the number of sales. This is equal to sales times the rise in the utility of the marginal consumer. However, the value to society is the integral of the rise in utility over all consumers. If the value of higher quality declines with the value put on the product as a whole, surely a reasonable hypothesis, the monopolist will undervalue and underprovide quality.

Spence suggests that "this aspect of market failure has very little to do with monopoly" but "is, rather, a result of the fact that price signals carry marginal information". However, it is at the very least much more acute under monopoly since a competitive market would be likely to generate supply at different qualities, which would attract intramarginal consumers.

Monopolists undersupply quantity as well as quality but regulation is intended to stop that happening. A price is set which ensures that the optimal demand is forthcoming and met. But quality still causes a problem. Vickers & Yarrow (1988) explain in a diagram (fig.1) also reproduced in Rovizzi & Thompson (1992) and Cave (1993). Suppose an increase in quality which moves the demand curve D(p,q₁) to D(p,q₂). It will increase revenue by p times the number of extra sales induced by the quality increase. Where demand is relatively price inelastic, or marginal consumers value quality less than the average, and where price discrimination is not possible or allowed, this is likely to be smaller than the value of the increase in quality to all consumers. The area S will exceed P . Quality will be undersupplied. This is seen as a defect of RPI-X regulation. Littlechild (1983) warns against the effect of a tariff reduction scheme on quality and says that it would seem sensible to ensure that quality of service did not deteriorate by requiring the regulator to monitor it and act if necessary.

Rate of return regulation is not seen as presenting such a problem. Spence hypothesises a case where capital (k) is the only input so the cost of producing x units at quality q can be written as a return on capital c(x,q) = rk(x,q). The rate of return constraint is pD/k £ s or

p £ (s/r) . (C(D,q)/D) .

Call the average cost function, C(D,q)/D, a(D,q). Differentiating the rate of return constraint as an equation, i.e. assuming it is biting, gives dp/dq= (a_DD_q + a_q)/(r/s - D_pa_D). If average costs rise with quantity (or, at any rate, do not fall too fast which would, of course, create other incentives towards quality), this is positive so the monopolist is able to raise price when quality increases which in this respect is an improvement on the position under RPI-X. The higher the allowed cost of capital (s) and the lower the real one (r), the more price can increase with quality.

Of course the dichotomy between the two methods is artificial. A rate of return method sets prices for a year, and sometimes longer. RPI-X sets prices (about) every five years with reference to a cost of capital or allowed rate of return. MMC reports and regulators' price control announcements have regularly based their conclusions on a rate of return or present value calculation. Thus, RPI-X might be described as rate of return regulation with a five-year time period and rate of return as one year RPI-X.

The previous section concluded that the incentive to invest in quality improvement was in part a function of the ratio of allowed to actual cost of capital. If a company cannot earn its cost of capital, it will resist investment. If the allowed return exceeds the cost of capital there will be an incentive to "gold plating". The cost of capital is not the subject of this paper. Recent price control reviews and MMC reports have appeared to reach a consensus on a pre-tax real return of around 7%. Regulated businesses initially argued that the cost of capital was much higher than 7% but the figure seems high as a cost of capital for a monopoly utility business (as opposed to a company engaging in monopoly and other businesses) and there is interesting debate as to why UK and US experience appears to differ. The 1998-99 price control round in energy, water, and rail may yet reach different conclusions.

Rate of return and RPI-X approaches do have important differences. The length of time before there is a rate of return adjustment clearly has an incentive effect. It is this which produces a realisable profit from increasing efficiency and cutting costs. The higher that incentive, the greater the inducement is likely to be to avoid, or at least delay, quality improvements. Costs of improving standards of service which are not related to capital investment are deterred in much the same way as efficiency is encouraged since, unless the improvement in output can be measured, they will feature purely as costs.

The means by which assets are added to the rate base also affects the incentive to provide quality if the regulated monopoly runs the risk of making an investment which is not subsequently included in the base. This might occur because the regulator considered the investment to be inefficient or the gain to consumers insufficient to warrant it or as the result of a "yardstick calculation". In the 1993 electricity supply price control proposals Professor Littlechild allowed each REC £1m p.a. in IT development expenditure when their expenditure and proposed spending varied widely. OFFER developed an equation using kWh supplied and customer numbers and mix to explain REC operating costs and based allowed revenue on the results. A company which spent more on IT or supply business staff in order to increase quality would not have been able to recoup the expenditure but one that reduced spending and quality (and was not detected in doing so) would increase profits.

A similar approach was taken in the distribution price control review the following year. An estimate of operating costs and the services of non-operational capital stock was made in part from regression analysis. Operational capital spending was benchmarked against the quantity of electricity distributed and the age of the existing capital stock. This would have had a similar quality incentive effect and so the review also "made an additional allowance for improving quality". The second distribution price control proposal publication also contained a discussion of the monitoring of proposed investment spending and the difficulty of reconciling the aims of ensuring quality improvements are made and of encouraging efficiency gains. This is addressed in the next section.

In those utilities' businesses where it is not possible to introduce significant competition, regulators have responsibility for setting the appropriate price/output mix. Spence suggested that the level of quality provided might be set as a by-product of that decision. "Ideally the regulatory authority would manage price-quality trade-offs by confronting the firm, on behalf of consumers, with a reaction function that reflects rates of substitution between price and quality on the demand side of the market. But these rates of substitution are difficult to determine, because computing them requires knowledge of the value attached to quality by the full range of intramarginal customers. Rate of return regulation has some merit in these circumstances as a second-best strategy."

UK regulators have not accepted this. It is not clear that rate of return would provide a good approximation on quality and the incentive properties of RPI-X have been preferred. This has required regulators to make explicit judgements about the quality of service on behalf of customers but they have not generally found it easy.

Rovizzi and Thompson list four main approaches. These are discussed briefly and then brief comments are made about their application in some UK regulated industries.

There are no pure monopolies. Every product has a substitute. Even in the regulated industries rail competes with air, sea and road and telecommunications can obviate the need to travel. Electricity and gas compete with oil, coal, insulation and clothing. However, television, telecommunications, postal services, rail, electricity, gas, and water are all regulated as possessing a degree of monopoly power.

The regulator sets minimum standards which have to be achieved by the licensee. Annexes to the licences state minimum times per week (averaged over the year) for different types of programmes and annual performance reviews record what has been achieved as well as commenting on other issues including making subjective assessments of programme quality.

Revenue is not decided by the regulatory process and so only the first and last approaches are available. Both are used but setting the minimum standards is bound to be a subjective process and important dimensions of quality cannot be easily measured.

As in television the degree of monopoly is declining. Statistics of performance and customer satisfaction are compiled and published regularly and some compensation can be claimed by customers for low quality. Quality aspects, such as network congestion, have been of importance in competitive parts of the market and competition is likely to be the main determinant of service quality in the future.

Service frequency and reliability are the main quality variables. These are not separately priced. Statistics are published and levels appear to be set by the company in consultation with the state as owner-cum-regulator. It may be that postal services represent the closest approximation to a monopolist (albeit facing competition from other communications) subject to a rate of return constraint and taking decisions on quality mainly in the light of that. However, there is significant ability to provide services of different quality and price and to discriminate between classes of customer.

The letting of each train operating franchise involves the Franchise Director acting on behalf of customers and taxpayers and taking judgements on the quality of service offered, subsidies required, and pricing restrictions. Performance statistics are published, passengers are compensated for poor performance, and allowable revenue can be reduced for poor performance. Thus all four methods are used to some extent.

However, a major determinant of the quality of rail services is investment undertaken by Railtrack, which provides the infrastrucure. The Rail Regulator in 1997 put in place a licence amendment which allows him to take enforcement action if the company fails to produce reasonable plans or to adhere to them without good reason. He therefore joins the electricity, gas , and water regulators in taking explicit judgements on the adequacy of the monopoly network provider's investment plans and subsequent expenditure.

Output targets are now being established for performance. The December 1998 ORR consultation paper says that the conclusions of the price control review "will incorporate commitments by Railtrack as to the performance, quality and capability of the network and the stations on it ...... [which] will be in the form of measurable and monitorable targets .... whose delivery is supported by an appropriate incentive structure".

Electricity companies conform to "guaranteed" and "overall" service standards. The former are undertakings to each customer and compensation is paid in the event of failure. The latter are percentage targets to be achieved (e.g. supplies reconnected within a given time after an interruption). They are subject to force majeure clauses, particularly severe weather, whose interpretation has caused some disagreement.

The main network performance indicator is the number and length of interruptions. These are published by OFFER and companies are required to publish Quality of Supply Reports. The review of investment plans at the time of price controls sets an implicit target but it is not published and monitored (or even calculated) because of the variable and unmodelled impact of the weather and other factors. The regulator is then placed in the difficult position of monitoring expenditure and deciding whether the normal (and sometimes large) subsequent underspend against plans represents an efficiency gain or a failure to supply the quality which had been promised. Scottish HydroElectric did set a target for the reduction in minutes lost of supply and OFFER's 1995 proposals suggested that other companies might do the same so that performance could be more easily monitored. Companies have since stated their targets in their quality reports.

When they reviewed Scottish HydroElectric's price control proposals the Monopolies and Mergers Commission recommended more investment to improve the quality of supply to remote customers than the Director General had suggested. The value of lost supply implicit in the recommendation appears to be higher than that used in the England and Wales Electricity Pool to determine the degree of interruption to electricity supply as a result of a shortage of generation. That value was itself set artificially high in order to preserve pre-Vesting security standards. High implicit values for lost load are not unusual in such schemes and illustrate the difficulty in arriving at an efficient solution when quality is decided by an essentially political process and involves a cross-subsidy from other customers when the possibly more efficient option of cash compensation might be less politically acceptable.

In the revision of distribution price controls for 1995 OFFER considered linking quality of supply and performance but "concluded that the difficulties and disadvantages of making an explicit link outweighed the potential advantages". This conclusion is presently being reviewed but without much expectation of change. This may in part be because only 2% of customers in OFFER's 1997 survey expressed any form of dissatisfaction with the quality of their supply. These are likely to have been in comparatively remote (and so probably expensive and cross-subsidised) areas. An economically efficient solution might result in a reduction in quality.

The gas industry has service standards like those in electricity and the Network Code, by which Transco contracts with gas shippers, includes a liabilities package intended to produce appropriate quality incentives. OFGAS is consulting about developing it. The linkage with the price control is difficult because of uncertainty about the appropriate level of investment etc. and the funding of compensation payments which would be made when the economically efficient level of service was provided.

Projecting and monitoring capital spending created significant problems during the last price control review and was the subject of disagreement between British Gas and the regulator who had commissioned reports on an underspend from W.S.Atkins. The MMC said that investment should not be reviewed within the price control period but that "in some circumstances it may be more appropriate to provide for recovery of any under-investment in the next review but in such a way as to maintain the incentive to Transco to minimise the cost of its capital investment (for example, by not recovering all the under-investment)". The advice highlights the difficulty of resolving the problem.

The problems of assessing and funding quality have been recognised longer and more explicitly in the water industry then elsewhere. When the industry was reorganised and privatised in 1990 a system was introduced whereby companies submitted plans to the regulator which were scrutinised and monitored by his staff and by appointed "certifiers".

OFWAT commissioned a willingness to pay study to help in setting standards but concluded that other factors must be taken into account. Obligations were imposed from outside which might conflict with the results. There were problems of affordability, which raise interesting questions of efficiency and cross-subsidy but, unlike the Scottish HydroElectric example, also significant externalities. Thirdly, the findings were approximate and exemplified some of the general problems enumerated by Cave:

• altruism - for example customer surveys regularly reveal a desire for special services for the elderly not wanted (at least for themselves) by those actually in the age group.

However, the Director General of Water Services is considering linking company revenue to performance. Under revised price controls.

Quality of service is becoming a more prominent feature of monopoly regulation. Increasing competition is enabling quality decisions to be taken in the market place to a greater extent. Where that does not happen regulatory intervention is inevitable. Such intervention is difficult to manage and regulators are constrained when deciding that reduced quality would be preferable to higher prices, particularly when groups of customers would be affected disproportionately.

A market mechanism would remove a great deal of the burden. It can be implemented through either compensation payments or the price control. Problems of valuation remain and customer surveys have not been successful in solving them. However valuation is implicit under any system of determining quality and it could be made easier by being explicit and refined over time (even though there would still be distributional effects which would make the values contentious). Uncertainty of revenue resulting from weather effects could be addressed in the insurance markets.

11. J.Tirole discusses this in "The Theory of Industrial Organization" MIT 1988.