On Markets and Regulation
by John Williamson, The World Bank
Paper presented to a conference held at the University of California, Santa Cruz
November 20-21, 1998
This paper was written while Mr. Williamson was the Chief Economist for the South Asia Region at the World Bank.
© Institute for International Economics
In most markets competition provides a better way than regulation of determining how much of a good will be produced and at what price. In most markets, therefore, the government should concentrate its policy attention on ensuring that the market is truly competitive, with regulations limited to those general rules that define the institutional infrastructure of a market economy and any specific rules that may be needed to ensure that buyers are aware of the quality of what they are purchasing or that the social interest in causes like a healthy environment is respected.
In the first part of this paper, I will consider the policies that India should adopt in the markets for most goods and services, those where a competitive regime is potentially possible. I then discuss the case of the "natural monopolies", where regulation to set maximum prices is desirable. A third section discusses whether the need for regulations to circumscribe the workings of the market or to set prices may best be administered through public ownership. A fourth section treats the factor markets, while the final section turns to consider the financial system.
Goods and Services where Competition is Feasible
Crude liberalizers have sometimes given the impression that all that needs to be done in order to establish a market economy is to privatize and abolish regulations that impede competition. This is far from the case, as the difficulties of the transition in most of the former Soviet Union bear witness. A market economy requires an institutional infrastructure, consisting at a minimum of a system of contract law, corporate law, and bankruptcy law, together with a judicial system that will enforce those laws impartially, rapidly, and efficiently.
Unlike the FSU when the transition began, India already has the main elements of this institutional infrastructure. Perhaps the two major respects in which the infrastructure is still deficient concern bankruptcy law and the speed with which the legal system enforces contracts.
In India a firm is supposed to seek permission from the Bureau for Industrial and Financial Restructuring before it closes down. T.N. Srinivasan (1998) reports that in the 11 years this Bureau has existed it has registered 2145 cases and "disposed of" just 220 of them. Presumably the remainder are still "sick" industrial units, either receiving subsidies to continue operating or having stopped operating but being prohibited from selling their assets such as land and buildings, thus preventing some other firm that could use them profitably from doing so. Apparently many firms are happy enough remaining in this limbo, because they have easy access to credit or because their managers may be able to engage in quiet piecemeal asset-stripping for their own benefit. There is obvious waste and inequity involved in such a regime, as opposed to a bankruptcy code like Chapter 11 in the United States or its equivalent in the other industrial countries, designed to avoid asset-stripping for the benefit of the managers and to allow assets to continue to be operated, normally by a different set of managers (although under specific circumstances the law may allow the original managers to continue in office, without, of course, continuing subsidies).
The other great weakness of the institutional infrastructure is the delay that is prone to occur before the courts act to enforce contracts. This is certainly a problem to the banking system in recovering payments due on bank loans, but presumably the difficulties are wider than this. In some parts of the country the problem of delay in enforcing contracts is reinforced by a fear of trying to enforce contracts, which unsurprisingly results in a higher level of non-performing assets (Rangarajan 1998).
In addition to these general issues, the operations of individual industries are almost inevitably subject to regulations specific to that industry, intended to ensure that firms operate with due regard for concerns like safety and the environment. Regulations intended to ensure safety may sometimes be rationalized by the criterion suggested earlier, of ensuring that individuals are well-informed of the likely consequences of their actions, although in practice most of us are prepared to go somewhat beyond this and endorse regulations that compel people to do things that are good for their own safety even if this means over-riding individual choice (such as compelling the riders of two-wheelers to wear safety helmets). Environmental regulations provide the classic case of attempts to internalize externalities.
Two criteria would seem to be relevant in judging such regulations. One criterion is that regulations should be cast in the form of general rules that do not require the exercise of administrative discretion, so as to minimize the opportunities for rent-seeking. The other criterion is that rules be limited to those subjects where a case can be made that some externality exists or that individuals could not be expected to judge quality or safety for themselves (or do not give it a proper weight in their decisions).
Beyond that, there is no case for seeking to regulate markets where competition can be brought to operate. The policy question is then how to introduce and maintain competition. Once upon a time India used to maintain a whole web of controls and restrictions that impeded competition, including industrial licensing, price controls, capital issues control, and sundry forms of protection against imports and domestic production by foreign firms. Most of this "license raj" has been dismantled since 1991, to the regret of almost no one except presumably the bureaucrats who used to get rents for issuing the licenses.
The main surviving remnant of the license raj is the system of small-scale industry reservation, which prohibits firms above a certain size (with an investment above Rs. 30 million) from operating in more than 500 industries, unless they export more than 50 percent of their output. Industries affected include garments, several forms of oilseed processing, toys, hand tools, and food processing. The small firms are not able to deliver large volumes of consistent quality, which has been a major constraint to the export of garments. SSI reservation along with protection resulted in market fragmentation, with the inefficiency of the reserved sectors passed on to consumers in terms of high prices and to primary producers in terms of low raw material prices for such products as oilseeds and cotton. It also resulted in the construction of excess capacity. In the case of oilseeds, for example, capacity utilization ranges from 10 percent for the smallest units to 30 percent for small-scale expellers whose counterparts in developed countries achieve a capacity utilization rate of 70 percent. Low capacity utilization along with the use of inefficient technology raises unit cost of production significantly compared with international norms: it costs 40 percent more to process a ton of soybeans in India than in China, and 90 percent more than in the United States. The intention was to promote labor-intensive operations, but the price that is paid for any success that the system may have in accomplishing that objective, in terms of stunting the growth of promising enterprises and curtailing exports of the labor-intensive products in which India has a comparative advantage, would seem prohibitive. The system is also riddled with anomalies, such as the fact that Indian firms can avoid the constraint by relocating to Nepal and foreign firms are not subject to the constraint if they locate in India. The only reasonable policy would seem to be to abolish small-scale industry reservation entirely.
Abolishing deliberate constraints on domestic entry and competition is only a first step. In the tradable goods industries, the next steps are to open up the economy to imports and to allow foreign firms to establish production facilities if they so wish, by abolishing restrictions on FDI. Policy has recently moved toward welcoming FDI, but the progress toward easing import restrictions seems to have stalled, despite the continuing presence of quantitative restrictions on the import of many consumer goods (which is a real anomaly in the late 1990s). It is difficult to understand why the Indian government seems to regard a slow phase-out of these restrictions as desirable: most economists recommend the immediate abolition of QRs, if necessary by replacing them with equivalent tariffs, a switch that leaves the domestic industry with much the same level of protection but that brings revenue to the government instead of rents to the protected industry and that allows the level of imports to adjust in response to economic incentives. The reduction in tariff protection may be done gradually, so as to give firms accustomed to protection a chance to adjust to foreign competition, and it may need to be accompanied by a devaluation, if that proves necessary to avoid the average firm becoming less competitive. It need not necessarily go all the way to unilateral free trade, for there may be a respectable case for maintaining a bargaining counter for trade negotiations, and a large country like India might even gain some national advantage through maintaining a modest incentive to broaden the industrial base. But allowing foreign competition is almost always the most effective way to introduce real competition into the domestic market, and India still too often fails to take advantage of it.2
Not all goods and services are tradable, however. To prevent competition in these sectors from being suppressed, it is desirable to have a competition (anti-trust) policy, which gives some quasi-judicial body the authority to investigate cases of alleged collusion and price-fixing and to penalize firms found guilty of abuses. It is not necessary to have any formal injunction preventing this body from dealing with cases that might involve tradable goods, it is just that one expects that in an open economy most cases of abuse of market power will involve the non-traded sector. India lacks an effective competition law.
It has been well-understood for many years that there are some industries where competition could not be expected to work because they are natural monopolies. The extreme case is that where there are increasing returns to scale throughout the output range up to where one firm supplies the whole of the natural market (which will be greater than the national market in the case of tradable goods, assuming open markets). In that case the largest firm has an advantage over all the others, and this advantage increases the larger the firm grows, so that competition will enable it to drive all the others out of business. Preventing this happening would be inefficient.
The standard policy recommendation in the case of a natural monopoly is price regulation. The regulatory body should be independent of the government, but it must apply principles that are given to it by the political process. Those principles involve a conflict between two objectives: that of limiting the profits of the monopoly firm to a normal level, and that of giving the monopolist an incentive to produce at least cost. The reason that the two objectives conflict is that the natural way to pursue the first one is to follow a cost-plus pricing rule, but that rule results perversely in the monopolist getting larger profits the larger its costs are. Regulators seek ways around this dilemma by searching for measures of what costs should be rather than what they are, such as "x percent less than last year" or what they are in some other country or region.
But how many industries are really "natural monopolies"? At one time we tended to put all the utilities into this category: electric power, gas, railways, roads, buses, airways, telecoms, ports, water supply, sewerage. Nowadays technology has made a number of these potentially competitive. Power generation can be separated from transmission and distribution, and power can then be traded on a competitive market, with the generators competing against one another. Railways compete against road transport over most of their loads, with only a few areas in which the natural advantage of railways is great enough to make regulation necessary. Buses and airlines can certainly be required to compete. Cellular telephones have introduced competition into telecoms, although regulation still appears desirable for basic services. Interestingly, however, it transpires that the more important role for regulation in telecoms is that of governing access between networks, which is essential if competition is to be able to function. Port services would seem to be an activity in which most cargoes could go by alternative routes, which makes it not obvious that continued regulation is necessary. Water supply and sewerage remain textbook cases of natural monopolies, so far as I am aware.
Thus the cases in which regulation would seem to be essential are quite limited: the sale price of electric power to the public, the margin to be charged for the transmission of power, the price of gas to the public and perhaps to domestic suppliers, some rail services, the rules governing access between telecommunications networks, and water supply and sewerage. But a number of other activities that were traditionally thought of as part of monopolistic industries can be opened up to competition: power generation, buses, airlines, cellular phone services, and ports. These activities, other than cellular phone services, remain without competition, and largely in the public sector, in India.
The Question of Public Ownership
That raises the question: when does public ownership make sense? This is a question on which many of us have revised our opinions since Mrs. Thatcher first began privatizing things in Britain in 1980. At that time I certainly subscribed to the view that this was only an important issue when privatization offered the opportunity to introduce the stimulus and discipline of competition, and that in other cases the enterprise would in any event be run by managers who would not be owners and who would therefore have much the same incentive to do a professional job of running their company whether they were reporting to shareholders or the government. The improvements in performance subsequent to privatization that have been witnessed in many industries even when their market position has not changed makes this position untenable. It seems that a better model than the Berle and Means separation of ownership and control is the one which says that capitalist businessmen have an incentive to maximize profits because otherwise they will be taken over while bureaucrats-as-businessmen have an incentive to maximize the size of their bureau (read workforce) because there is no threat of takeover and size is what determines their salaries and prestige (and/or an easy life).
That gives a general advantage to private ownership. Are there any countervailing factors that could outweigh that consideration in particular instances? There may be, in cases where regulation is necessary for any of the reasons developed above, i.e. either because of the need to take account of externalities, or because competition is not practically possible, or (as in telecoms) because regulation is actually necessary for competition to prevail. Satisfactory regulations need to be able to spell out the external considerations that should be taken into account and how they must be taken into account. And satisfactory price regulation needs some way of estimating how low costs potentially could be. If one could get the right people running nationalized industries, who could be relied upon to take external considerations into account in an appropriate way and to minimize costs because of professional pride, it is conceivable that public ownership could offer a preferable way of achieving the objectives of regulation. Water in England is a case where I am still unconvinced that privatization made sense.
But it is quite clear that vast areas of the Indian economy are in public ownership that cannot conceivably be rationalized by this type of logic. In 1992 4.7 percent of industrial units with almost 55 percent of the installed capital, 27 percent of employment, 23 percent of gross output, and 26 percent of value-added, were in the public sector, and since then there has been little disinvestment of majority ownership. All the tradable goods industries, including coal and steel, as well as electricity generation and distribution,3 airlines, telecoms, and ports seem unambiguously ripe for privatization.
This does not mean that I am now an indiscriminate advocate of privatization. In fact, I would argue that India has gone too far in one instance, in seeking to persuade the private sector to build toll roads. Rural roads are, after all, the classic case of a public good, where consumption is non-rival and only with inconvenience and added cost excludable. Consumption of urban road space is of course highly rival, especially during the rush hour, but the pricing of urban road space that economic efficiency demands would surely have to be done by the public sector (it certainly is in Singapore, the one place where it has already been introduced). The idea of having the private sector build a particular stretch of road in return for the right to levy a toll on it is totally different, for the effect is to drive traffic from the safe, fast, uncongested new road on to the slow, dangerous, congested but free old road.4 This is exactly what has happened in Pakistan, where a toll is being charged on the spanking new M2 motorway that has kept much traffic on the parallel and congested Great Trunk Road.
Most factor markets are far from being competitive in India. It is difficult to fire people because of the law, even if the employer no longer has work for them to do, in firms with more than 20 employees. Pay in the public sector is determined from time to time by the Pay Commission, apparently with no reference to the balance of supply and demand; the latest pay award has aggravated the fiscal crisis, especially in the states.5 The Urban Land Ceiling Act prevents changes in the pattern of land use, keeping land unnaturally scarce in city centers by forbidding its sale to any firm in a different sector.
Presumably these policies are intended to prevent the exploitation of the weak by the powerful. There are surely instances in which one can sympathize with market interventions that have that objective; I imagine, for example, that most of us would be capable of sympathizing with laws that limited or maybe even forbade child labor,6 because the person making the decision as to whether the child works is not the child itself but a parent that may not be maximizing its welfare, or even if it is the child one might question whether the child has enough information to make a sound judgment of the costs and benefits. Indeed, prohibition of child labor is one of the four core labor standards that the ILO seeks to secure in all countries, along with freedom of association (the right to unionize and the right to collective bargaining), freedom from forced labor, and freedom from discrimination in employment and pay.
But in other cases a terribly high price is paid for these good intentions. For example, the evidence seems to show fairly conclusively that firing restrictions cost jobs, because employers are so reluctant to run the risk of needing to pay surplus labor that they simply avoid taking on so many people in the first place (Fallon and Lucas, 1991). It is even more difficult to think of any rationalization for the overpayment of most of the civil service, or for prohibiting changes in land use. One piece of good news is that the government is now seeking the repeal of the Urban Land Ceiling Act.
The Financial Sector
Financial markets differ from most other markets in several fundamental ways. They are inherently intertemporal: they involve lending today in the expectation of a return in the future, which introduces an element of risk that is usually largely absent in non-financial transactions. It is likely that the borrower will be better informed about the probability that he will be in a position to service his debt in the future than the lender: this is the problem of asymmetric information. This may result in adverse selection, a situation in which those most willing to borrow are not those with the best probability of being able to service their debts. And it may be that the lender will be less worried about this than he should be because of the problem of moral hazard, meaning that s/he succumbs to the temptation to rely on an explicit or implicit guarantee of some third party rather than make those loans that promise the best risk-return combination from the borrower. Moral hazard may also distort the borrower's decisions, for example if s/he believes that there is no need to service the debt because the courts will procrastinate.
The inherent characteristics of financial contracts noted in the previous paragraph suggest that one should expect the liberalization of the financial sector to be more complex than that of most goods markets. In particular, I think that the attempt to allow for these factors means that we must expect to see financial markets being subjected to regulations that have no parallel elsewhere in a market economy. Let me outline four areas in which regulation seems to me defensible.
1. Prudential supervision of banks is already widespread, and is increasingly being standardized internationally to the norms laid down by the Basle Committee. Note, however, that what recently counted as best practice in terms of bank supervision has already been superceded by a new approach in which bank examiners examine a bank's risk monitoring system rather than a snapshot view of its balance sheet. And the crisis that started last year in East Asia, most particularly the misfortunes of Long-Term Capital Management, has now raised concerns that best-practice risk-monitoring systems have so far failed to make adequate allowance for the covariation of important risks that results from the phenomenon of contagion.
2. It seems desirable that a degree of prudential supervision of nonbank financial intermediaries (including hedge funds) should follow before long. The principles will have to be discussed, but I would think that some minimum capital adequacy standards and some requirements for transparency are likely to feature among any requirements.
3. My colleague Joseph Stiglitz (1993) has argued that the bank deposit interest rate should not be allowed to exceed the interest rate on short-term government paper. His logic is that this would prevent banks "gambling for resurrection" by offering high interest rates to depositors confident that a government guarantee will secure their repayment even if the bank fails. The UTI debacle provides a graphic illustration of what happens when this precept is ignored. The asset side of UTI's balance sheet would suggest that it should have been required to act as a mutual fund, and not allowed to offer a guaranteed rate of return above the government interest rate, a guarantee that now seems likely to need a government bailout if it is to be honored.
4. The current crisis has suggested to many observers that precipitate liberalization of the capital account is a costly error. It follows that it would be inappropriate to incorporate capital account convertibility in the Articles of the IMF, which should instead encourage countries confronted by excessive capital inflows in the future to take the same sort of prudent measures that Chile and Colombia did.7 India already seems to have recognized that the timetable suggested by the Tarrapore Committee is now unrealistic, and will for the time being keep its limits to the foreign exchange exposure of both the bank and corporate sectors.
I am not for a moment suggesting that financial liberalization has already run its course. But the next reforms in this sector might focus on privatization rather than further liberalization, for surely private bankers would not tolerate the risk aversion that has now paralyzed lending to the corporate sector in India, any more than they would feel comfortable with the risk courting practiced by UTI. Likewise insurance is a sector much in need of the stimulus of competition.
The East Asian crisis has posed three threats to the Indian economy. The most immediate one, the threat of financial contagion, did not materialize, because India did not have the risk exposure to the international market that was present in Indonesia or Korea or Russia. The threat from a weakening of the current account induced by world recession is still developing, but it seems likely to lead to a slowdown rather than a catastrophe. The third threat is in the realm of ideas: the danger that people will draw the wrong lessons from the crisis, and interpret it as casting doubt on the proposition that in most areas India needs more market discipline and competition, and less regulation. But this is a proposition that we can preach with conviction only if we also recognize the limits of its validity, and are clear about the specific areas in which regulation is needed if a market economy is to function successfully.
Fallon, Peter R., and Robert E.B. Lucas (1991), "The Impact of Changes in Job Security Regulations in India and Zimbabwe", World Bank Economic Review, vol. 5(3).
Rangarajan, Indira (1998), "NPA Variations Across Indian Commercial Banks: Some
Findings", mimeo (NIPFP).
Srinivasan, T.N. (1998), "Eight Lectures on Indian Economic Policy", mimeo (Yale).
Stiglitz, Joseph E. (1993), "The Role of the State in Financial Markets", ABCDE, World Bank.
1. The author is indebted to conference participants for useful comments on an earlier draft, and to Forhad Shilpi for research assistance.
2. For example, just before the conference Surjit Bhalla (The Economic Times, 10 November 1998) noted that the Indian Oil Corporation, itself a public sector monopoly, had done a deal with Reliance that made the latter sole provider of a range of products and prohibited their import except under extreme conditions of non-availability.
3. Power distribution can be broken up into sufficiently small and similar units to allow a regulator to use their relative performance as the basis for determining their sale prices.
4. I will admit two qualifications. First, bridges may have so few substitution possibilities as to be exempt from this critique. Second, the objections would fall away if the private sector's revenue were based on a shadow toll rather than an actual toll.
5. Admittedly this is in part because the government implemented only the pay-raising part of the award, and not the efficiency-raising part that was supposed to help pay for it.
6. The case for hesitating to forbid child labor is the fear that under certain circumstances children can be forced into much worse occupations, like child prostitution.
7. Both countries adopted a requirement that those taking any loan from abroad should make a non-interest bearing deposit equal to some proportion of the loan with the central bank for a fixed period. This reserve requirement was more expensive the shorter the term of the loan, resulting in a deterrent against taking the sort of short-term loans that were the vehicle of contagion in East Asia.