Methods of privatization.

Berg, Andrew ; Berg, Elliot

Privatization has been in fashion for more than 15 years, if we date its recent flowering from Margaret Thatcher's initiatives in the late 1970s. Shrinking the state's economic presence became part of the economic reform programs that characterized economic policy throughout much of the world in the 1980s. The collapse of communism in Eastern Europe and the former Soviet Union (FSU) moved privatization issues onto a much larger stage in the 1990s.

Privatization in the broadest sense means giving private actors a greater role in decisions about what, where, and how to produce goods and services. A great deal of experience has now accumulated regarding this process. Some of it shows the great potential that privatization has for increasing productivity, income and welfare.(1)

However, it also reveals the complexity and difficulty of effective privatization. Privatization even in the narrower sense of "divestiture"--the sale of state-owned enterprises (SOEs)--has presented greater challenges than its early advocates envisaged.

Relatively few countries account for most of the divestiture activity in recent years. Thus, of the roughly $300 billion in divestitures between 1988 and 1994, two-thirds occurred in industrial countries. (This excludes sales of small enterprises and give-aways such as "mass privatization" programs.) The five biggest privatizations--among them Japanese Railways, British Telecom, and Deutsche Telekom--account for some $60 billion in proceeds. Of the $100 billion in developing country privatizations, more than half took place in Latin America.(2)

In developing countries, macroeconomic impacts of privatization activity have been small in most cases. Despite the widespread privatization rhetoric and many formal structural adjustment programs under World Bank sponsorship, the average public enterprise sector share in gross domestic product for 40 low- and medium-income developing countries between the late 1970s and the early 1990s remained unchanged, as did the sector's share in wage employment.(3)

In the low-income developing countries (those with average per capita incomes below $600 a year) the pace of privatization has been particularly slow. In sub-Saharan Africa, for example, only 1800 (mostly small) divestiture transactions took place between 1980 and 1995, with a sale value of about $2 billion. In only six countries (Benin, Ghana, Guinea, Mozambique, Nigeria, South Africa and Uganda) did total sales through 1995 amount to more than $50 million, which suggests insignificant reductions in the relative size of the public enterprise sectors in most of the continent. Ghana, Nigeria and South Africa account for almost 70 percent of the total sales value of African privatizations.(4)

Overall progress has been much faster in Eastern Europe and the FSU. Most of these countries have succeeded in moving from almost complete domination by the state sector to predominantly private economies.(5) However, only a few of these countries have been able to divest many state enterprises to new managers, and the problem of using former state assets more efficiently remains problematic in much of the region.

Many factors enter into the explanation of these patterns of privatization. In this paper, we focus on only one of these: the relationship between methods or techniques of privatization and the objectives that governments seek to attain through privatization. This has two related dimensions: the fit between method and objectives, and the incompatibilities between various objectives.

The main objectives, explicit or implicit in most privatization programs, are: fiscal relief by cutting government subsidies to money-losing SOEs and/or by generating new revenues from their sale; increased enterprise efficiency; increased efficiency of the entire economy through more competitive markets and better allocation of resources across firms and sectors; increased political support and broadened institutional underpinnings for a market-based economy or further liberalization; stronger financial markets; increased investment and the stimulation of entrepreneurship.

The main methods to be reviewed here fall into five broad categories: sales of shares or assets; capital dilution; management-employee buy-outs (MEBOs); broad-based or mass privatization; and indirect or partial privatizations via management contracts, leases or service contracts. These categories are convenient but not entirely mutually exclusive: MEBOs are a type of sale, while mass privatization involves some combination of the other methods with some type of free distribution of shares or vouchers.

Before we turn to a discussion of these techniques, we should note that we are leaving out a number of privatization instruments, the most important being privatization through liberalization.(6) The creation and growth of new private firms, along with a withering of state enterprises, has been even more important in many of the transition countries than privatization in terms of the reduction of the state's ownership share.

Consider two otherwise dissimilar examples: Poland and China. Poland's private sector has grown from 29 percent of gross domestic product (GDP) in 1989, largely in agriculture, to 56 percent in 1994. Though classification and data problems make the calculation difficult, most private sector growth in Poland appears to have come from the creation and growth of new firms, not from privatization of existing firms.(7) The workers and, in some cases, the assets for these private firms presumably originated in the state sector, but these were pushed out by economic pressures on state enterprises or pulled out by the attraction of high profits in the private sector. China is perhaps an even more dramatic example, though definitional and measurement issues are even more difficult. The share of the non-state sector, including so-called township-and-village enterprises (TVEs), in gross industrial output rose from 24 percent in 1980 to 57 percent in 1994, solely through differential growth rates following liberalization.(8)

This is clearly privatization in the broadest sense, but a full discussion of this method lies outside the scope of this paper. Suffice it to say that macroeconomic stability, price and trade liberalization, elimination of restrictions on start-up firms and a real hardening of the budget constraints facing state enterprises appear to be required to allow fast privatization, as well, perhaps, as a tolerance for the methods by which assets find themselves in private hands.(9)

Turning now to privatization itself, we consider each of the five methods in turn. We describe each briefly, provide relevant examples, indicate prevalence, summarize strengths and weaknesses and show how various methods entail trade-offs between government objectives.

Sale of Shares or Assets

The classic type of privatization is the sale of full or partial ownership of a state enterprise by public offering on stock exchanges, by competitive bidding for shares or assets or by non-competitive placement of shares.

Public Flotation of Shares

Under this method, the state sells to the general public through the stock market and other financial institutions all or a substantial part of the stock it holds in a going concern. The initial public offering (IPO) is often combined with other methods, such as the sale of shares to employees on favorable terms.

The public flotation is politically appealing and has great revenue-raising potential. It allows broad ownership, which is always more popular than a sale to powerful domestic or foreign buyers. Wider stock ownership is a common objective in most privatization programs, as it was for example in the United Kingdom, Jamaica, Chile and more recently in Germany. It also has the effect of locking in privatization actions. Most observers believe, for example, that renationalization of Chile's telecom SOE is unlikely because, as a result of mandated preferences for small investors, one-third of the shares of the major telecommunications company has passed to the general public.

Public flotation is also flexible. It allows targeting of particular groups to meet political objectives or social purposes. Thus, in some Jamaican privatizations, as in Chile and elsewhere, small buyers were given preference. In the sale of the Jamaican National Commercial Bank, for example, no investor was allowed to own more than 7.5 percent of the outstanding shares.(10) Sale via public flotations can also contribute strongly to the development of local capital markets, as in Jamaica, where the initial privatizations increased the capitalization of traded shares by 40 percent.

Public offerings are also more transparent than other methods. Prices are set by the market for all to see, and for anyone to buy. Since one of the main obstacles to privatization is widespread public concern about corruption and cronyism, this is no small advantage. Sale through the stock market can be accompanied by private placements--to pension funds, to disadvantaged groups or to employees, thereby increasing the equity of the transaction.

In addition, public offerings allow gradual approaches. Sales can be organized in tranches, with first tranche prices and conditions being designed to win acceptance, and later tranches set at higher prices to benefit from improved enterprise performance and better knowledge in the market about the privatized firm.

The characteristics of IPOs that make them attractive also make them hard to implement, especially if speed is an objective. Their clarity and transparency bring tremendous transaction costs--such as preparation for sale, valuations and managing the offer. As a result, only larger SOEs or large government holdings are usually appropriate. Moreover, firms have to be readied for sale. They must be made attractive to buyers. At the same time, the diffuse ownership that tends to result means management will likely not change, nor will owners exert strong control after privatization. This often means restructuring, government absorption of debt and new investment. Also, market conditions have to be right. Where markets are not buoyant, offerings can fail. This happened recently in Indonesia, Brazil, Turkey and some other countries.(11)

Conflicts of objectives are inherent in setting the offer price of shares to be sold. Governments may seek a high price to achieve its revenue objective and to avoid later charges of giving away crown jewels at fire-sale prices. The objectives of winning political and market acceptability, however, dictate a low price. But pricing shares too low not only invites later political attack, but erodes another objective of widespread popular ownership. Low-income buyers and employees tend to sell their holdings if share prices rise rapidly after initial offerings.

The case of Polish bank privatization illustrates the potential importance of this pricing problem. The government determined that it would privatize the major commercial banks partly through IPOs, with the rest sold to employees or to a strategic investor or remaining with the Treasury. After one successful offering in April 1993, the second bank sale in December of that year turned out to be substantially underpriced, with shares skyrocketing more than 13 times on the first day of trading. The large windfall gain this provided to those who were able to buy the shares initially, and in particular to the employees and management who were able to purchase 10 percent, led to public outcry. Partly for this reason, the program was halted so that it could be reformulated. No further major banks have been privatized in Poland, despite forecasts in 1994 that the remaining seven large banks would be privatized before 1997.

These pricing dilemmas and how they have been handled have generated some disappointment with privatization in Europe. A recent article in the Economist laments the fact that, while privatization has broadened financial markets (since 1985 it has added about 1 percent a year to the capitalization of listed shares in European stock markets), it has not been successful in turning Europe into a continent of shareholders. While 20 percent of U.S. households own stocks, and around 15 percent in Britain, the comparable figure for Germany and France is still not much more than 5 percent. One reason is that small investors have not always done well. In France, privatized company shares have underperformed the market; a consistent buyer of privatized shares since 1990 would in fact have lost money.(12)

There is also a conflict between providing financing for employee purchase of shares (installment sales) and the risk that share prices will fall, leaving employees with net losses and debt obligations. This can be avoided by providing repurchase guarantees--buy-back arrangements that insure workers against losses. In recent flotations in Spain, Italy and Germany, all investors have been sheltered from falling share prices by such devices. But such protective measures also entail a conflict. As the Economist noted: "Such tricks seem to subvert the point of privatization. What purpose, after all, is served by teaching potential investors that owning equities is all profit and no risk?"(13)

Public offerings result in diffuse ownership and give no assurance that assets will be better-managed than under government ownership. Where SOEs have been operating in competitive markets and/or autonomously, and hence have not been notably inefficient, this may not matter. This was apparently the case with the privatization of the National Commercial Bank in Jamaica, where the goal of increased efficiency at the enterprise level was secondary or absent; political acceptability through diffuse ownership mattered much more. However, sacrifice of efficiency objectives undermines a major rationale for some privatization, especially in poorer and slower-growing countries. Many governments therefore try to combine public offerings with sale of substantial ownership shares to core or strategic investors who can bring about restructuring through better management, new investment and access to new markets.

One privatization story from Bangladesh illustrates how stock market flotations of minority ownership shares can fail to improve management, while sale of majority ownership to a technical partner can lead to vastly better performance.(14) The stateowned Kohinoor Chemical Company, a major producer of soaps and cosmetics, sold 49 percent of its shares in a public offering in 1988. Going public did not affect corporate performance, however. Because of poor management and marketing, annual sales fell by two-thirds between 1989 and 1992. The labor force grew from 350 to 1200, and overtime payments also grew. A widely cited example of management laxity was the employment of 27 drivers for the company's fleet of 5 cars; drivers and security staff nonetheless claimed substantial overtime.

In July 1993, the Bangladesh government's 51 percent share in the company was sold to the highest of four bidders. Soon after, the work force was cut to 650 people. Overtime disappeared. Production processes were overhauled and new marketing efforts undertaken. Part of the terms of sale was that the company would retain the 650-person work force for a year. Since only 350 employees were needed, the management introduced a scheme of team competition. The workers were divided into two teams, each working alternative weeks. Output per worker has shot up--five-fold, according to some observers. Production has increased by 30 percent. The company has offered continued employment to all its workers based on future profitability and expansion.

Given the advantages and difficulties of public offerings, it is not surprising that they are found principally in the developed world and among the more advanced of the developing countries. Indeed, most industrialized country privatizations have been through public offerings. In low-income and transition economies that have not had well-developed stock markets, financial institutions or regulatory arrangements, IPOs have been much more infrequent.(15) They are not unknown in the more developed transition economies, especially in Hungary, the Czech Republic and Poland.(16) In sub-Saharan Africa, there appear to have been fewer than 70 public flotations, in only five countries, namely Cote d'Ivoire, Ghana, Kenya, Nigeria and Zambia.(17)

Variants of public offerings have been devised in conjunction with mass-privatization approaches. Mass forms of IPOs have been rare for the same reasons that it is difficult to arrange even one. Some countries that distributed vouchers widely allowed their use for the purchase of shares in individual enterprises, most commonly in the form of auctions, such as in Russia, the Czech Republic and Mongolia. In Mongolia and in the former Czechoslovakia, large enterprises were sold through centralized voucher auctions, with the resulting shares trading on a stock exchange.

Competitive Bidding

Sale of either part or all of an SOE's shares or assets by public tender is the most common privatization instrument worldwide. Most small firms in the transition economies were privatized by auction, as were many firms in the developing countries. For small retail shops, small-scale transport and service operations generally, auctions are quick, present few valuation problems and can generate revenue for the state budget. Problems of asymmetric information, when present managers know more about the firm than outsiders, are relatively small. In any case, the markets in question are easily contestable; competition is likely to prevail.

Competitive bidding is a component of most mass-privatization schemes implemented in recent years in Eastern Europe and the FSU. In Russia, most medium and large enterprises have been privatized in part through public auctions of shares for vouchers (coupons), which had been distributed nationwide for a nominal fee. There, and in Mongolia, the vouchers can be used to buy shares of privatized companies. In the Czech Republic and Slovakia, investment funds bid for shares in thousands of medium and large state enterprises in a carefully structured, multi-step nationwide auction.

Sales of medium- and large-scale enterprises and, in poor countries, even smaller firms are usually done by formal tendering. As with share flotations, transaction costs can be substantial, if the tendering is done right. The accounts of the firms to be sold have to be brought up-to-date and audited, outstanding liabilities and asset values determined and future profitability estimated. Pre-privatization analyses should yield suggested minimum selling prices and should result in the preparation of an information memorandum for potential bidders. Governments are frequently interested in other terms of sale besides price--whether the transaction will be for cash, for example, and, more importantly, whether the buyer commits to maintaining employment or injecting new investment.

For larger firms, the pretransaction preparatory work is often done by investment banks or management-consulting firms, and not infrequently with financing from foreign aid sources. The consultants also help with marketing through contacting potential buyers and with advertising. Tender announcements appear weekly in the Wall Street Journal, the Economist, the International Herald Tribune, and similar publications.(18)

Sometimes assets are sold rather than shares in entire going concerns. This may be done because state entities are too small to justify the costs of being corporatized--being put in joint stock company legal form--before privatization. But other reasons are more important. The government may want to spin off unrelated or unprofitable divisions or subsidiaries before selling the core enterprise. Sale of assets allows this kind of unbundling. It also may be the only way to deal with complicated situations. In the case of a Nicaraguan construction company, regional branches were sold to workers and management as separate units; heavy equipment at the main site was transferred to the state public works department, while the main office was returned to the former owners.(19)

Some companies cannot be privatized by share sales (in other words, as going concerns) because uncertainties about their contingent liabilities are too great. In this case they can be dissolved and liquidated and their assets sold without attached liabilities. A new company (very often with the same name) then rises from the ashes. This is an extremely common phenomenon, as well a major source of confusion in statistics on numbers of privatizations and liquidations.(20)

Sale of SOEs through competitive bidding avoids the major deficiency of public flotations: uncertain impact on corporate governance and therefore on improved firm-level efficiency. Most of these transactions entail sales of going concerns; they are of the type commonly called "trade sales." Buyers are technical partners or core investors who will take over management and will have incentives to enhance profitability. This presumably means cost-cutting, increased productivity through better organization and new investment and a search for new market opportunities. The objectives of increased investment and penetration of new markets are especially likely to be achieved through sales to foreign investors.

The objective of raising revenue may also be achieved by competitive bidding. However, the proceeds will depend on whether restructuring costs and debt obligations have to be assumed by the central government. Proceeds depend also on how wide the net is cast in seeking buyers. Purchase can be restricted to nationals in order to meet the political acceptability goal, but this will reduce the fiscal impact. At least as important, and often overlooked, is the fact that proceeds depend on the intensity and effectiveness of marketing. Too often in the past, marketing efforts have been perfunctory, with few ads appearing in the international press. Few potential buyers appear, and the selling price suffers accordingly.

Tradeoffs often emerge in the bidding process itself. Political acceptability demands objective criteria and complete transparency--such as the listing of bid prices and the publishing of bases of all awards. This works in the direction of acceptance of the highest bid. But subjective elements invariably cloud the decision. Some bidders are more credible than others, and most governments have multiple selection criteria, such as whether injections of new investment will be made, debt obligations absorbed and commitments about employment entered into. In the end, bids are often not comparable. Thus, even where well-specified procedures are carefully followed and the process is wholly above-board, competitive bidding may not dissipate all suspicions of foul play. It is nonetheless usually the most even-handed and transparent option available.

The most pervasive trade-off in the privatization process concerns time--the sacrifice of key objectives for speedy action. A recent study of experience in privatizing telephone companies in Argentina, Jamaica, Mexico and Venezuela illustrates the point.(21) In some of these countries, haste and a desire for revenues crowded out other goals and led to deficient regulatory systems and concession agreements that were, in many instances, unfavorable to the privatizing governments.(22)

In Argentina, the original plan had been to privatize with a view to increasing competition. It had envisaged breaking up the national telephone monopoly into two regional companies; free entry into long-distance services was to be allowed immediately, with entry into all services allowed after only five years. All cross-subsidies (financing low price domestic services by charging high prices for long distance) were to be abandoned. However, as negotiation unfolded, these ideas were largely dropped to get a quicker sale. The goal of generating revenues, and demonstrating to investors that the economy was truly opening up assumed higher priority. Very low targets were set for expansion and quality, and post-privatization regulatory rules were sketchy.

The cost of haste is most evident in the case of Jamaica, which was pressed to meet World Bank conditions regarding the amount of revenue derived from SOE sales. Buyers received a 24-year monopoly (compared to 6 years in Mexico), and no specific targets were set for expansion of service, nor for quality of service; penalties for non-performance were also inadequate. The Jamaican buyers were guaranteed a real, after-tax annual rate of return of 17.5 to 20 percent.

More privatization transactions are completed globally by competitive bidding than by any other method. In industrialized countries, which mainly use public flotations, it is not the most common method but does occur. Examples include the sale of Leyland Bus and Truck in the United Kingdom in 1986, and more recent sales of water and power companies. It is probably the most common method in middle-income developing countries. In the transition economies, competitive bidding using vouchers was used to buy shares in Mongolia's first wave of privatization. Russia's sales also were by competitive bidding, using vouchers. In both cases, however, there are problems of definition, since insiders received such preference that the process resembled management-employee buyouts more than competitive bidding.

Trade sale-type bidding (one firm bidding to buy control of an SOE) has been infrequent in the transition economies. Only in Estonia and Hungary have privatizations by sale to outsiders (non-employees or managers) been substantial.

Data on transactions by technique applied are also available for sub-Saharan Africa.(23) Out of some 1800 transactions between 1980 and mid-1995, almost half were by competitive bids for all or part of going concerns. It has been the most frequently-used technique in that part of the world.

Noncompetitive Sales or Transfers

These kinds of sales take many forms. The state may find it advisable to sell to a preselected buyer without competitive bidding. This is sometimes called a private placement. A good strategic investor may have made an offer that meets the government's price and other requirements, and officials may decide therefore that further bidding is superfluous. Some or most of the shares (or assets) may change hands in this kind of transaction, which may be accompanied by public flotation, award of preferential shares to employees, or other placements.

A noncompetitive sale may follow a failed tender. In Guinea, five bidders responded to a 1995 request for bids to purchase a controlling share of the government telecommunications monopoly. After careful assessment, all these bids were rejected. A later proposal by the Malaysian telecommunication company was accepted for want of a better alternative.(24)

Placement of shares with insurance companies, pension funds and other financial institutions is a common noncompetitive means of privatizing ownership. Restitution, the return of companies to former owners, is of course noncompetitive. Joint ventures or mixed companies are other noncompetitive methods. Private partners often have so-called preemptive rights, the right of first offer when the public partner decides to sell shares. This is a common situation where private owners retain some ownership after bouts of nationalization, as for example in Zambia, where recent privatization actions included numerous sales to those with preemptive rights. A key problem here is determination of the sales price of shares, since there are no market values to go by.

Finally, transfers of shares to trusts can be included in this category of methods. Various kinds of transactions are currently in operation. The most famous outside the transition economies is the special trust set up by the Malaysian government as part of its preferential program for native Malays, the Bumiputra. Shares of public corporations reserved for the Bumiputra are deposited in this trust, which now holds a significant share of the total capital stock in the country.

The Zambian Privatization Trust Fund is a variation on this theme.(25) It holds in storage, for future sale to small local investors, shares in privatized companies. Part of the rationale is to distance the government from the newly privatized activities; government no longer votes the transferred shares. The rationale also is that the fund's managers can operate without haste and in an environment that is better informed than at the time of implementation of the privatization program. The fund can issue shares through public offerings and by sales to pension funds, insurance companies, and other intermediaries. It can give small investors special incentives and can set down ownership limits to avoid undue concentration. The board of the fund is drawn from the private sector and management is contracted by competitive bidding. If shares are not sold after five years, the fund will become a unit trust and will be sold to Zambian investors. Unsold shares could be given away free to Zambian citizens.

Mass-privatization programs have also involved a transfer of sales to trusts. In Poland, the government established National Investment Funds (NIFs) and distributed shares from about 500 medium and large enterprises to these NIFs. Each citizen was then allowed to purchase a certificate of ownership (voucher), which was essentially a share in each of the NIFs, for a nominal fee.

The benefit of noncompetitive sales is that they can be cheaper, easier and quicker than alternative methods. Particular technical partners with special competence can be sought out, creating good prospects for more efficient management. Negotiations can be more flexible than those in formal bids. Political and social objectives can be well-targeted; for example, shares can be distributed to underprivileged groups, or to employees and other stakeholders or to insurance companies and pension funds. Entrepreneurship-nurturing objectives can be served. The trusts may be able to serve some corporate governance role.

Noncompetitive approaches thus satisfy many objectives. Aside from the allocation to trusts, however, these noncompetitive approaches suffer from one major disability: they are inherently less transparent than competitive methods, and such transactions are often attacked as unfair or corrupt. Negotiation with one potential buyer also often entails long and frustrating dealings with slippery partners.(26) Perhaps for these reasons, noncompetitive bidding in privatization is not very prevalent in most countries. A sub-Saharan African inventory suggests that about a third of the 1,800 transactions (1980-1995) that involved sales of shares or assets were done noncompetitively.(27)

Capital Dilution or Capitalization

Privatization can occur without the state disposing of any of its equity, but rather adding to it by allowing a private investor to buy in. The result is a capital increase, with the government's share declining. Many joint ventures are formed this way. In Nicaragua, the government put up the routes of its Aeronica Airline for a 51 percent share in a joint venture, with the private sector partner contributing cash, equipment and management. Capital dilution is also an easy vehicle to transform partially private companies to more fully private firms, as in the case of the 1995 capital injection to the Guinea Industrial and Commercial Bank, which reduced government's equity from 89 percent to 12 percent of the total.

Capital dilution/joint ventures are often politically acceptable, since the government retains a large or even majority share in ownership. If the government is a passive owner, allowing full autonomy to the private management, economic efficiency goals are likely to be well-served. This approach gives undercapitalized enterprises new life, although problems of working capital scarcity may remain. It is also fast, especially where existing partial ownership by a core investor is supplemented by new capital injections. However, risks do exist regarding transparency and equity, notably in setting share prices. In many countries there are no rules for how share prices should be determined in cases of internal acquisition. Using book values in prosperous enterprises (especially banks) can result in an underpricing of the government's shares.

The Bolivian capitalization scheme is a variant of capital dilution. The government is currently transferring to private hands half the ownership and all the management of the six biggest public enterprises--the rail, air, power, phone and petroleum distribution monopolies, and mining smelters. International investors invest new capital for up to half the equity of the enterprise. Core investors sign management contracts, including an option to purchase additional shares later. The new capital can be used for investment or working capital. The remaining half of the equity will be held by 5 to 10 new private pension funds, which will provide retirement and disability benefits to shareholders.(28)

The main disadvantage of the Bolivian approach is that, by giving away its interests in SOEs, the government may be sacrificing the revenue-generating objective. This loss of revenue may be mitigated, however, to the extent that the transfer of assets to pension funds substitutes for other resources the government would have had to put into the pension funds. The delays caused by the need to work out the details of the new pension scheme is a minor inconvenience. On the other hand, powerful advantages are evident. The removal of government from ownership of the SOEs deepens the credibility of the government's commitment to privatization and market reform. This increases the attractiveness of the firms to be sold, especially to foreign buyers. The privatized entity benefits from the availability of new resources for working capital or investment.

In addition, although not directly due to the capitalization scheme, Bolivia has been able to adopt a near state-of-the-art strategy for privatizing of its major public utilities involving the abandonment of old notions about "natural monopolies." In the case of electric power, for example, monopolies can be broken into at least three components: generation of power, transmission and distribution to users. The idea is to introduce competition on the generation and distribution sides while giving equal access to the transmission facility. This approach to privatizing infrastructure has been pioneered in the United Kingdom, New Zealand, Chile, Argentina and the United States. Bolivia now joins the front ranks of countries that are privatizing infrastructure with an emphasis on competition. The government describes its program thus:

The Ministry of Capitalization of Bolivia intends

to transfer the electricity transmission network in

the Bolivian interconnected system, currently

owned by the state company ENDE, to the private

sector through an international public bidding to

be completed in the first quarter of 1997. The transfer

of the transmission network to the private sector

is the latest stage in the wholesale restructuring

of the Bolivian electricity system. This process has

involved the passing of a new electricity law and

regulations, the introduction of a new regulatory

structure for the industry, the capitalization of three

new electricity generating companies formed from

the generation assets of ENDE (through the subscription

of 50 percent share holdings in the capitalized

companies and granting of management

control to three leading US electricity operators)

and the sale of two major distribution companies

to Spanish and Chilean utilities.(29)

Africa is the only region for which there is data regarding the incidence of privatization by joint ventures or capitalization. The African inventory data to which we have referred above records only 34 joint ventures or capital dilution privatizations between 1980 and 1995, out of more than 1800 transactions.

Management-Employee Buyouts (MEBOs)

Three main types of Management-Employee Buyouts (MEBOs) can be distinguished. In many developing and transition economies, small establishments--for example, retail outlets, restaurants, hotels, bookstores, trucks and buses--are sold to employees. Countless privatizations have taken place this way, though they do not seem to show up in privatization inventories. One reason is that they are often under $50,000. Another is that they may be listed as liquidations, since state enterprises are often legally dissolved before the assets pass into employee hands.

In Eastern Europe and the former Soviet Union, tens of thousands of small establishments were privatized by transfer to employees.(30) In Guinea and Mozambique, dozens of regional state food distribution outlets became employee-owned or leased. In Nicaragua, buyouts were encouraged by a policy that set 25 percent of each privatized firm aside for employees. The assets of one gold mine in that country were sold to workers employed anywhere in the mining industry; in the construction company fragmentation, six regional entities were sold to employees.(31) The World Bank's recent study on African privatization was able to identify only 22 MEBOs, about one percent of total privatization transactions. This is certainly very much below the true figures, for reasons suggested above.

The second type of MEBO involves employee stock ownership of medium- and large-scale enterprises. These are fairly common in industrial countries. The first major privatization of this type was in the United Kingdom with the sale of the National Freight Corporation in 1982 to a consortium of existing and retired employees and 4 banks, with employees taking over 80 percent of the shares. Examples are also found in developing countries. In Chile, employees bought most of the stock of Soquimich, a sizeable producer of nitrates. When ENDESA, the state-owned power company in Chile, was broken up for privatization, one of the units, EMEL, was sold to employees. There have been cases in Asia and elsewhere, though usually of middle-sized companies.

A recent example from Pakistan gives the flavor of these transactions.(32) In January 1992, employees of Millat Tractor Limited took over management after purchase of 51 percent of the shares. Most employees participated, financing 40 percent of the purchase with cash and the rest with a bank loan secured by their pension funds. They won out over four other bidders. The employees elected seven of the nine directors, although five of the seven are nominated by the lending banks. About half of the 49 percent nonemployee share is held by stockholders who purchased shares through the stock exchange, and the other half is held by government financial institutions. Of the 51 percent worker share, about 30 percent is held by workers and 70 percent by executives. The company has made some difficult decisions since and appears to be thriving.(33)

The third and numerically most significant form of MEBO is de facto insider domination of nominally open privatizations Examples in transition economies abound. Most of Russian industry has been privatized this way. The Russian model took the form of a voucher program that gave special preferences to employees of state enterprises, usually resulting in MEBOs in practice. In Mongolia, workers (and their families) chose to buy shares largely in "their" enterprises, ending up with 45 percent of the total shares, according to one survey.(34) Most enterprises privatized in Poland have been privatized through so-called liquidation.(35) This has taken various forms, but involves the sale of either the enterprise or its assets through auction for cash. In most of these, the managers and/or the workers ended up owning controlling shares after privatization.(36)

MEBOs address one of the central obstacles to privatization, particularly in the transition economies where ownership rights and traditions are vague and not well-protected: the fact that finding new owners means disenfranchising existing stakeholders. Enterprise "insiders," such as workers and managers, have a traditional and de facto claim on the enterprise. Which insiders are important varies from country to country: in Russia and Hungary, it was the enterprise managers and government bureaucrats who ran the enterprises prior to the fall of the Soviet Union; in Poland, where this group was partly disenfranchised by Solidarity, it was also workers. Insider power derives from a variety of sources. One is political; in some countries, managers and/or workers have political power and can control the privatization process, or at least block it. Another source of power is legal; in Poland, workers voted on privatization schemes for "their" enterprises. More generally, though, insiders both care more about "their" enterprise than most others and know more, which gives them bargaining leverage. Especially where court systems are weak and laws nonexistent or incomplete, de facto control can be vital. In Russia, enterprise managers have at times simply locked shareholders out of meetings or not informed them of the time and place.

MEBOs of shops and other small establishments can be speedily and easily done. There is no faster way to establish unambiguous private property rights over state assets. They are also relatively equitable and favorable to employees and thus may be politically palatable. However, they fail the revenue generation test, with prices sometimes so low as to approach free distribution. MEBOs also have a mixed score on the entrepreneurship-encouragement criterion. On the one hand, evidence from several studies suggests privatization that leaves the original management in charge results in much less entrepreneurial activity than other methods.(37) On the other hand, entry of new investors over time is presumably facilitated through the establishment of property rights, and in any case the counterfactual may be no privatization, not privatization to outsiders.

MEBOs are the vehicle of choice for small companies in general and for doubtfully saleable enterprises in particular. Efficiency effects should be positive because of privatized ownership, although management competence may not change, except if outsiders are brought in later. Moreover, insiders may have important information advantages over other owners. Established industries with stable technologies and low capital-labor ratios, which are producing in competitive markets, are most likely to have a successful MEBO. In any event, opportunity costs (sacrificed revenues and imported management) in these circumstances are likely to be small, and few economic distortions will result.

For larger enterprises, MEBOs are generally more problematic. They can entail obligations by the government to take over enterprise debts. Sales often have to be financed on credit of one variety or another, and that can lead to drawn-out involvement, since many MEBOs of this kind do not succeed, and governments are often induced to take them over again. The evidence on the impact of MEBOs on enterprise restructuring is not yet clear. In one survey of Polish enterprises, employee-owned firms restructured much more aggressively than state-owned or outsider-owned firms, but this may have been due to the fact that insiders chose to buy the more promising enterprises.(38)

Mass Privatization

Mass privatization programs privatize hundreds or thousands of enterprises at one time. To make this possible, they generally combine one or more of the above techniques with some sort of free distribution of assets, shares or buying power over assets (vouchers). Something along these lines has been used or is under consideration in a number of transition economies, including the former Czechoslovakia, Russia, Poland, Romania, Lithuania, Kazakhstan, Kyrgyzstan and Mongolia. It is the subject of widening discussion as an option in other developing countries, such as Tanzania and Uganda.

It is easy to see why this approach is appealing. In the transition economies, almost all industrial production and most output in other sectors was generated in the state sector. The huge scale of privatization alone ruled out the "classical" sale as a single or main method; it would take decades, except in the East German case where vast subsidies and willing buyers allowed the rapid privatization of some 8,000 enterprises.

There were other reasons for the appeal of this form of privatization. Valuation of state enterprises in these economies was difficult and uncertain, given the lack of a market track record and persistent distortions that ruled out easy estimates of existing viability and future profitability. Also, small domestic savings, weak or nonexistent capital markets and embryonic market institutions made private buyers few and wary. Foreign investment could and did play some role, but the new governments and their publics found unacceptable the idea of selling off large parts of the economy to foreigners.

In sorting through these various programs, it may be useful to consider that each program has two components: a "supply side" which dictates how enterprises will be chosen and prepared for privatization and what will be sold, and a "demand side" which determines how ownership in the privatized enterprises will be allocated.(39)

Each of these "sides" presents a number of key issues. On the supply side, the preparation of large numbers of enterprises for privatization in a short amount of time can be extremely difficult. For example, the legal definition of the firm often must be determined--i.e., does it include shared assets such as a municipal power plant? More importantly, the claims of various existing stakeholders such as workers and managers must be dealt with. As described in the section above on MEBOs, these insiders often have substantial de facto or de jure power over "their" SOE. These aspects of the problem are related; preparation of enterprises for privatization on a large scale requires the cooperation of the insiders.

On the demand side, the problem is how to find owners for hundreds or thousands of enterprises rapidly and in ways that promote good enterprise governance, equity, public support and capital market development. A fundamental question is how widespread distribution of shares or vouchers can translate into effective governance of the privatized enterprise; when ownership is extremely diffuse, no one owner has much incentive to monitor or discipline management. More generally, the question remains as to whether the pattern that emerges will be conducive to capital market development, including a well-functioning stock market.

To illustrate these issues, we will briefly describe the mass-privatization programs in Poland, the Czech Republic, Russia and Mongolia.

The Polish program, proposed in 1991 but implemented only in 1996, cover approximately 500 medium and large enterprises. Enterprise participation in the program required an affirmative vote by the workers. Workers receive a relatively small share of the privatized firm--in effect, they receive 10 percent of the shares free of charge--and, while they anticipated some benefits from privatization, they also feared its effects. As a result, it has not been easy for the government to convince large numbers of promising enterprises to participate, delaying the program's implementation.

On the demand side, the designers of the Polish program wanted widespread and fair distribution of ownership in the enterprises, but they were concerned that diffuse ownership would fail to provide the strong owners required to supervise difficult but necessary enterprise restructuring. They therefore transferred most enterprise shares free of charge to 15 National Investment Funds (NIFs), to act as mutual funds of sorts, owning most of the shares of each privatized enterprise and in turn owned by the population. The NIFs are managed by private fund management companies, often with the participation of foreign financial institutions that were selected by the government and given incentives to raise the value of their holdings.

One-third of the shares of each company went as a block to one NIF, ensuring that each company had one owner with a substantial interest and hence incentive to provide good corporate governance. Each citizen was allowed to purchase a voucher, essentially a share in each of the NIFs, for a nominal fee. The initial Boards of Directors of the NIFs were appointed by the government, allowing some political control over the NIFs.

It is too early to judge how effective this program will be. The program met with stiff political resistance, partly from enterprise insiders and partly from those fearing the power of the NIFs. Although its design was essentially complete in 1991, its implementation was delayed until 1995. Certificates became available for purchase in November of 1995 for 20 zlotys (about $7 now) and, by the end of the subscription period in November 1995, certificates had been purchased by 26 million people. With a current secondary market price of about 150 zlotys, the government is encouraged and reportedly is considering further rounds.

The Czech Republic also distributed ownership widely, in this case through the sale of vouchers at nominal prices to each citizen. In Czechoslovakia in 1992, all adult citizens could buy for about $30 a voucher booklet allowing them to purchase 1,000 shares in up to 10 companies. They could do this either through one of the many investment funds set up by banks and others, or could bid directly for the shares.

On the "supply side," the powerful central government had weak insiders to contend with as a result of a system that remained highly centralized until the end of 1989. Thus, the enterprises had no choice as to which firms the government would identify to participate in the program. For each enterprise, the government chose a privatization project from among those submitted to it, and anyone could submit a project. Reflecting the insider advantage even in Czechoslovakia, the projects that were eventually chosen in most cases had been submitted by the managers of the firm.

These projects specified how the shares in the enterprise would be sold, whether through tender, voucher auction or cash purchase, among other options. In the first wave of privatization ending in December 1992, property worth more than $23 billion was sold, $7 billion of which was through vouchers.(40) Investment funds held approximately 70 percent of the shares sold in the first wave and a similar proportion of the voucher points invested in the second wave, which was completed at the end of 1994.

These two cases demonstrate the power of mass privatization schemes to privatize large numbers of enterprises quickly through free distribution and financial intermediaries. They also show, in quite different ways, the difficulties with such funds. Widespread ownership of investment funds can pose problems somewhat similar to the problems of diffuse ownership of enterprises themselves: who will control the behavior of the fund managers when their ownership is divided among thousands (or millions, in the case of Poland) of small shareholders? Poland chose a centralized approach, with the government creating the funds and managing the selection of fund managers. This proved politically difficult and, for many observers, raised the specter of government interference in the control of the firms after privatization. The Czechoslovakian approach avoided such a strong government role and resulted in a more spontaneous ownership structure.

It is too early, again, to say too much about how this emerging structure is working. On the one hand, there are some signs that funds may be helping to provide the impetus to restructure enterprises. On the other, the ownership of Czech industry and banking is largely concentrated in a few funds, which are themselves mostly managed by Czech banks. This raises dangers of conflict of interest between banks as lenders and banks as fund managers. It also raises the question of the government's ongoing role in these enterprises, as the government continues to own a large share of bank stocks. In recent months charges of lack of transparency in investment fund behavior, in particular with unregulated trading among investment funds, have spooked markets and perhaps discouraged some outside investors.

In the Russian case, the vouchers could also be used at auction, but the role of investment funds was more limited, and the structure of the auctions led to much more worker/management control of the privatized firm. The Russian mass privatization program ended in July 1994. It involved 100,000 medium-sized and large firms, including 15,000 in industry. More than 60 percent of the industrial labor force was affected. Each citizen born before 2 September 1992 (150 million people) received one voucher with a face value of 10,000 roubles. The vouchers were traceable and valid until mid- 1994, with the typical market price varying over time from 4,000 to 42,000 rubles. They could be used to buy shares in the enterprise where the owner worked, allocated to one of the many investment funds that emerged or used directly in voucher auctions. They could also be used to pay for housing and other property, and could also be bought and sold by anyone, including foreigners.

In order to get insiders to cooperate, the program was structured to provide them with substantial advantages. Employees had various options for buying shares, including the right to buy 51 percent in a closed subscription at a price of 1.7 times nominal value per share, paid for with vouchers and cash. This and other options gave employees and managers the possibility of acquiring majority voting stock and control of the firm. Typically, the legal minimum of company shares were allocated for distribution through public voucher auctions. Mass privatization became largely a matter of MEBOs, with more than two-thirds of privatized firms coming under insider control.

Not all vouchers went to purchase shares by employees in their privatized firm. According to surveys in mid- 1994, about a third of the voucher recipients had sold their vouchers and about 30 percent had exchanged vouchers for shares in voucher funds. It appears that about 10 percent of the privatized medium-sized and large firms in Russia followed the open access route.

In Russia, the voucher program was unquestionably successful at privatizing huge numbers of enterprises quickly. Despite insider dominance, privatization has achieved some of its broader aims.(41) First, massive depoliticization has taken place. Control rights have been transferred from government ministries to managers and investors. The sectoral ministries, only a few years ago the kingpins in Russian industry, have become irrelevant. The process of depoliticization is incomplete, since government still has many levers of influence such as credit policy, heavy regulatory structures, tax policy, and local governments have stepped in to control the system out of interest in social services. However, depoliticization has come a long way.(42) Secondly, Russia has made a giant step toward efficient ownership patterns. Outsiders have apparently been accumulating shares in the voucher and share markets. There are some signs of large shareholder activism, including removal of general directors at initial meetings.

The dominant role of insiders has nonetheless presented troubling features. First, restructuring has been slow.(43) Indeed, there seems to be little difference between the behavior of stateowned and worker- or manager-dominated privatized firms. In insider-dominated firms, managers have been unwilling or unable to carry out deep reforms, particularly in reducing overemployment. De nuovo private firms, in contrast, demonstrated much more dynamism. Second, capital market development has been hindered, with managers able to control the secondary market for shares and the behavior of boards of directors, thereby leaving little room for stock market discipline of poor managers.

Part of the reason for the poor performance of privatized firms may be that the benefits of privatizations come only after a time lag. Nonetheless, in the words of some of those involved in the design of the program:

The principal message we draw from our empirical

evidence is that restructuring requires new people,

who have new skills more suitable to a market

economy. A secondary message is that, without new

people, incentives for old people might not be particularly

effective in bringing about significant

change.... (C)ontinued control by old managers presents

a problem for restructuring, and...more attention

should have been paid to management turnover

as opposed to shareholder oversight over the

existing managers. To some extent, large investors

have begun to force old managers out...(and)old

managers have been given enough wealth that they

can afford to retire in peace and let a new generation

take over. This, however, is probably not

enough.... If privatization were designed from

scratch, these strategies should have received more

attention than they have...(44)

The Mongolian case is of special interest because Mongolia is both a transition and a developing economy. It is the only developing country that has used mass privatization/voucher methods. Coupons distributed in 1991 to each citizen were used in a small privatization phase to buy agricultural and other assets in auctions. In principle these were open auctions, but workers had first rights to make a joint bid. By early 1993, two years after the program was launched, all small urban businesses--and 90 percent of the national total--had been privatized. All agricultural co-ops and more than 90 percent of state farms were in private hands.

Privatization of Mongolia's large enterprises followed. Ownership in many of the large firms was transferred through auctions for vouchers. As with small privatization, these open auctions were ultimately dominated by the workers from the enterprise. Workers (and their families) chose to buy shares largely in "their" enterprises, ending up with 45 percent of the total shares, according to one survey.(45) However, changes in the way companies are managed have been slow in coming. The government retained control of many of the privatized entities, and preexisting management and staff has remained in place. Diffuse ownership has thus far failed to lead to shareholder consolidation via secondary trading because the larger denomination coupons are not traceable. The regulatory system surrounding capital markets remains sketchy.

Indirect Approaches

Management contracts privatize management, leaving ownership in state hands. Some involve straight fees. In most cases, however, payments are tied also to performance. The key issues in success or failure are whether performance is related to the contract terms, and whether managers have true autonomy in hiring and firing. A recent World Bank study found that management contracts have improved both productivity and profitability in most of the cases studied.(46)

Lease contracts are of different types, varying mainly by who is responsible for financing investment. Under straightforward leasing (sometimes called affermage) the contractor or lessee pays the public owner a fee for the right to operate a public facility and bears the financial risks of its operation. This method is widely used in power, ports, urban transport, waste disposal and industry. The private firm finances working capital and replaces nondurable capital assets, with contracts generally running 5 to 10 years. The contractors collect tariff revenues directly and pay a share to the government. Regulatory burdens are considerable.

The Russian "loans-for-shares" program of 1995 might be classified as a leasing approach. Under this system, private banks lent money to the government for one year, with shares of large and valuable natural resource-based SOEs serving as collateral. The banks bid for the shares at auctions, competing on the size of the loan they would make. If the government failed to repay the loan after one year, the banks could auction the shares for a "market price."

The government was evidently not up to the regulatory burdens implied by this program. It proved difficult to avoid collusion by the banks in the auctions, resulting in low prices. When the government failed to repay the loans, it also proved difficult to avoid insider dealing in the resulting auctions of shares. The program is widely considered to have led to the transfer of large quantities of valuable assets to a small number of private owners at low prices.

Concessions involve greater contractor responsibility than leases, notably for replacement of fixed assets. They also last longer--normally 15 to 30 years. Water supply, waste disposal, toll roads and ports are among the common areas of usage. One variant coming into wide use is the creation of two separate companies, a societe de patrimoine, which is an operating company that owns the assets and is 100 percent state-owned, and a societe d'exploitation, which is majority-owned by a private operator with minority state ownership. This approach has become popular in privatization of water supply and power facilities and state-owned plantations. It has registered some striking successes--notably urban water supply power in Cote d'Ivoire and water in Guinea. But it also raises problems of coordination, since the societe de patrimoine is responsible for new investment and may not always consult with the operating company about investment policies.

Contracting out (also known as outsourcing or subcontracting) is widespread in public-sector service provision. It is an extremely diverse form of privatization, especially common for municipal services, and is widespread in the United States. Examples include security and janitorial services, maintenance services, data processing, and food service.

The main objection to indirect privatization methods is that they can be and often are a reflection of government unwillingness to bite bullets, becoming a substitute for all-out privatization of ownership, or "real" privatization. However, management contracts, leases and service contracts can be first steps, opening up further privatization prospects in situations where full privatization is not feasible or desirable. The partial and tentative nature of these contracting arrangements can be beneficial in that it makes them more politically acceptable than full divestiture. In addition, asset valuation problems are reduced, because leases with low rental values are easier to swallow psychologically and politically than sales at deep discounts from book value. The arrangements are in any case temporary. They allow for joint learning and experimentation, easing the way to effective privatization later.

Implementation of management contracts and leases is not without problems. Too often they are on a fixed-fee basis, divorcing remuneration from performance. Often, also, governments are unable to deliver on their promises to keep hands off and/or to provide investment and other funds as specified in the contract. Capacity building is often neglected. Additionally, leases often suffer from ambiguity in specification of responsibilities, as in the societe d'exploitation--societe de patrimoine model mentioned earlier. This can lead to disconnects between operational needs and investment planning; the asset-owning SOE remains responsible for investments, a situation that demands close coordination, which is not always forthcoming.

Contracting out has the same advantages as management contracts and leases in terms of tentativeness, experimentation and political acceptability. Where workers' trade unions oppose outsourcing, its political acceptability is less complete. Efficiency impacts are especially likely, because it is easier to assure privatization into competitive markets. It also has the crucial advantage of being well-designed to meet the entrepreneurship-nurturing goals that should have highest priority in low-income countries.

Contracting out has other efficiency-raising advantages that are of special value in low-income country environments. It allows SOEs or central governments to employ specialized skills they cannot otherwise afford to recruit or retain because of low salaries in the public sector. It also allows use of specialized skills that are not needed full-time. It raises economy-wide efficiency by increasing specialization and by permitting greater exploitation of economies of scale and scope for specialized activities or functions, and creates yardsticks for cost comparisons with in-house operations. It tends, finally, to use less capital-intensive methods of service delivery and productions: it encourages taxis instead of big bus companies, corner traders rather than big state trading operators, decentralized maintenance shops rather than concentrated, large workshops.

Three main problems beset contracting-out privatization. First, it requires substantial decentralized capacity in bid preparation, specification of norms for contracted services and payment systems that are prompt. Second, it is open to abuse and corruption. And, finally, public employees, like private sector workers, generally dislike outsourcing.

None of this is excessively constraining, as the growth of subcontracting in many poor countries attests. But it does mean that special efforts are required to facilitate its wider use, and special precautions are essential to avoid corruption and favoritism.

Information on the extent of indirect privatization is incomplete and diffuse; it is not tracked as carefully as other types of privatization. A recent worldwide search came up with about 160 management contracts, a quarter of them in the hotel industry and another quarter in agriculture. More are found in Africa than elsewhere.(47) A survey in the early 1990s could identify only about 40 leases and concessions in infrastructure sectors, divided between power, transport, and water/sewerage and about the same number of contracting-out arrangements, almost all of them in transport and most in Africa.(48) However, these are surely underestimated.(49)

The best evidence for the incidence of contracting-out arrangements is in cities in the United States. Data shows that local governments are rapidly shifting to contracting out for many services.(50) The federal government is also a major user of contracting out. Its versatility is illustrated by John Glenn's comment after his space flight, that he could not keep from worrying a little bit when he reflected that his spaceship was built with components from 50,000 subcontractors.

Conclusion

No privatization glove fits all hands. A public offering may be right where the priority objectives are development of the stock market and wider spread of share ownership, and where government has adequate time and resources. Where overall economic growth is slow and public sector assets are poorly used, sale to an outside investor may be the most promising route; it is likely to lead to the fastest restructuring, especially in the case of trade sales to foreign investors. For the rapid sale of a large number of small- and medium-sized enterprises, auctions are the obvious method. Where speed is high on the priority list, as in transition economies with ambiguous property rights and weak controls where insiders have a strong stake, MEBOs are appropriate, in particular for small- and medium-sized firms. To privatize large numbers of enterprises rapidly, there is no practical alternative to mass privatization schemes involving widespread distribution of ownership. Where there is a very small supply of modern sector capitalists and entrepreneurs, methods that give the highest priority to entrepreneur-creation should move to the fore. Here small-scale opportunities have to be generated through the fragmentation of public sector activities and through the leasing and contracting out.

Much of the frustration that can surround privatization derives from the inability of any method to fully satisfy the multiple objectives that characterize typical programs. Public flotations generate revenues for the state, invigorate capital markets and broaden share ownership, but hasty implementation and poor price-setting may seriously dilute positive impacts for corporate governance. Trade sales, often the most efficient in terms of enterprise governance, raise nationalist hackles if foreign buyers appear, disturb egalitarians because of benefits to rich people and also may suffer from too-speedy implementation or insufficient transparency. Politically attractive and socially equitable instruments, such as mass distribution or transfers to trusts, often end up with insider-dominated buy-outs or concentrated ownership by investment funds; effects on productivity through better corporate governance are uncertain.

Two persistent trade-offs emerge--one between speed and the quality of privatization transactions, and the other between equity and improved corporate governance.(51) As the discussion of Latin American telecommunications privatizations indicated, pressure to implement quickly can often lead to relatively unfavorable terms for privatizing governments and to situations where regulatory arrangements are poorly defined. In the mass-privatization schemes of transition economies, from Mongolia to Russia where political conditions required insider cooperation, positive effects on enterprise productivity remain to be seen.

The analysis provokes two other general observations. First, it is easy to be simplistic in judging the "success" or "failure" of privatization programs. What is probably the most important criterion in most cases is the extent to which privatization sets the stage for future private-sector development--that is, whether it creates the political and/or institutional conditions that are favorable to further liberalization or marketization.

Judgements on this point are not simple. The approach of encouraging expansion around the state enterprise sector, as in China and Poland, has clearly led to extraordinary growth of the non-state sector, but the problem of inefficient public enterprises remains unresolved and is worsening. The Russian approach so far has not led to significant improvements in corporate governance, but whether it has blocked further change or made change easier is not yet clear.

An emphasis on laying the groundwork for future private sector growth implies a related conclusion. Policies regarding price and trade liberalization, demonopolization and deregulation, a hardening of budget constraints on SOEs, and regulatory systems that encourage external shareholder participation may be more critical for private sector development than privatization of ownership by itself.

Second, it is essential to prioritize objectives in defining privatization strategies. In many low-income countries, aid donors and local policymakers are now considering the adoption of mass privatization methods as a way to get lagging privatization efforts moving again. They may be right. But where, as in sub-Saharan Africa, the chief problem is the stimulation of entrepreneurship, it may be wiser to concentrate on methods that hit that target better, such as fragmentation or unbundling of functions, wider resort to leasing and contracting-out arrangements and creating an environment in which new private firms can thrive. (1) Ahmed Galal et al., Welfare Consequences of Selling Public Enterprises (Oxford: Oxford University Press, 1994)

(2) World Bank, Bureaucrats in Business. The Economics and Politics of Government Ownership. (Washington, DC: World Bank, 1995).

(3) ibid.

(4) World Bank, African Economic and Social Indicators (Washington, DC: World Bank, 1996).

(5) See European Bank for Reconstruction and Development, Transition Report, (London: European Bank for Reconstruction and Development, 1995).

(6) We also leave out privatization of finance (shifting payments for public goods and services from government budgets to private users) and many issues specific to privatization of infrastructure and the financial sector.

(7) Luca Barbone, Domenico Marchetti, and Sefano Paternostro, Structural Adjustment, Ownership Transformation, and Size in Polish Industry, (Washington, DC: World Bank, 1996).

(8) W. Tseng, E. K. Hoe, et al., "Economic Reform in China: A New Phase" (Washington, DC: International Monetary Fund, 1994); W.T. Woo, The Art of Reforming Centrally-Planned Economies: Comparing China, Poland and Russia," Journal of Comparative Economics (June 1994) pp. 276-308; W.T. Woo, Chinese Economic Growth: Sources and Prospects (University of California, 1996).

(9) This privatization through liberalization is clearly related to so-called "nomenklatura privatization" in some transition economies, in which enterprise insiders (especially managers and bureaucrats) appropriate physical assets or profit streams from the state enterprises they control. However, such practices are only part of the story; state enterprises will also release assets and workers if they are under enough financial pressure. See A. Berg "Does Macroeconomic Reform Cause Structural Adjustment? Lessons From Poland," Journal of Comparative Economics (1994) pp. 376-409.

(10) Roger Leeds, Privatization Through Public Offerings: Lessons From Two Jamaican Cases," Privatization and Control of State-Owned Enterprises, R. Ramamurti and R. Vernon, eds. (Washington, DC: Economic Development Institute of the World Bank, 1991) p. 98.

(11) Joseph Borgatti, "Methods of Privatization of State-Owned Enterprises," paper presented at the Methods and Practices of Privatization Conference (New York, 1993).

(12) "Privatisation in Europe: Is the Price Right?" Economist (23 November 1996) p. 87.

(13) ibid.

(14) M. Mohiuddin, Reform of the Bangladeshi Para-Public Sector: Situation and Prospects, (Clermont-Ferrand, France: CERDI, University of the Auvergne, 1996).

(15) It is also possible, of course, to use regional or even rich country stock markets. Ghana's Ashanti Goldfields Co. is listed in New York, for example, as are many privatized Latin American utilities.

(16) World Bank, From Plan to Market: World Development Report 1996 (Washington, DC: World Bank, 1996), p. 53 provides data on the incidence of privatizations in transition economies broken down by method.

(17) These typically were used only for small companies. Only 2 of the 20 biggest privatizations in sub-Saharan Africa as of the end of 1995 used public flotations (World Bank, 1996).

(18) In one week in October 1996, the International Herald Tribune advertised the sale of cotton ginneries in Uganda, and the Economist had an advertisement seeking qualified firms that could prequalify to bid for a 31 percent interest in Mozambique National Airlines. (19) Borgatti, p. 53

(20) There also may be tax reasons to sell assets rather than going concerns through share transfers--benefits from tax loss carryovers, for example.

(21) R. Ramamurti, ed., Privatizing Infrastructure in Developing Countries: Lessons from the Telecom and Transport Sectors in Latin America (Baltimore, MD: Johns Hopkins University Press, 1994).

(22) In Mexico, political and nationalist objectives led to stipulations that bidding consortia be controlled by Mexican nationals; the result was almost surely a lower price than might have been achieved. The other three countries allowed foreign control.

(23) World Bank data from a special study on privatization in sub-Saharan Africa (unpublished). Preliminary results are in World Bank, 1996.

(24) Helen Nankani, "Techniques of Privatization of State-owned Enterprises," World Bank Technical Paper no. 89, vol. 2 (Washington, DC: World Bank, 1988).

(25) World Bank, Private Sector Development in Low-Income Countries, Development in Practice (Washington, DC: World Bank, 1995).

(26) In Honduras, a steel-rolling mill that had never opened was sold first on a debt swap basis to a U.S. fabricator, who planned to use the plant to cast rail crossings. After a while this technical partner fell out, supposedly because it had been bought out by a French company that was not interested in the Honduras deal. The mill was then sold by debt swap to a Miami bank that was associated with the two original promoters of the first deal. The new group spent years searching for money and technical partners, while the mill remained unopened (Borgatti, p. 47). These kinds of problems are not restricted to noncompetitive sales; they also occur after formal tenders.

(27) Of 549 noncompetitive transactions, almost 300 were sales of shares, 135 were sales of assets, 77 entailed preemptive rights and 30 restitutions, while sales to trusts numbered 13.

(28) Example: A power company's assets are valued at $100 million. A buyer puts up $50 million in cash for 50 percent of the equity. The remaining state-owned 50 percent is given to pension funds. The new $50 million is available for company use.

(29) Advertisement by Bolvian authorities, International Herald Tribune (October 1996) p.8

(30) John S. Earle et al., "Small Privatization: The Transformation of Retail Trade and Consumer Services in the Czech Republic, Hungary, and Poland," CEU Privatization Reports vol. 3 (Budapest: Central European University Press, 1994).

(31) Borgatti, p. 49

(32) As described by Sharit K. Bhowmik, "Takeovers by Employees: A Response to Privatization in Pakistan," Economic and Political Weekly (29 April 1995) p. 931-933.

(33) The company had 830 employees when it was taken over, of whom a third were executives. Some 250 employees were quickly fired, with generous severance payments. Staffing in 1994 was 533, of whom 170 were executives. Labor productivity is said to have much increased; output has more than doubled; profitability has returned and the firm is diversifying its output and expanding.

(34) Georges Korsun and Peter Murrell, Ownership and Governance on the Morning After: The Initial Results of Privatization in Mongolia (College Park, MD: Center for Institutional Reform and the Informal Sector University of Maryland, 1994).

(35) Of the 3,075 enterprises privatized or in the process of being privatized from 1990 through 1994, 2,231 of them were done through "liquidation." See Central Statistical Office, Rocznik Statystyczny (Statistical Yearbook) (Warsaw: Central Statistical Office, 1995).

(36) Andrew, Berg, "The Logistics of Privatization in Poland," Jeffrey Sachs, Olivier Blanchard, and Ken Front eds. NBER Conference on Transition in Eastern Europe, (Chicago: University of Chicago Press, 1994). (37) Earle, page xxviii, says that they "have found clear evidence that the entry of new entrepreneurs, not connected with the predecessor retail establishments, is the most significant factor in increasing the levels of post-privatization investment." See also Barberis et. al.

(38) John S. Earle and Saul Estrin, "Employee Ownership in Transition," Froman Frydman, Cheryl W. Gray and Andrzej Rapaczynski, eds., Corporate Governance in Central Europe and Russia, (Budapest: Central European University Press, 1996) pp. 62-77.

(39) It is worth noting that "voucher privatization" is not a useful category, as it fails to specify what will be sold (the supply side) or even how the vouchers are to be used (MEBOs, auctions, IPOs and so on).

(40) Many of the best enterprises were apparently sold through tender, not through vouchers. In part, this reflects the fact that interested buyers and the enterprise management could negotiate a sale in the context of the proposed plan.

(41) See Maxim Boycko, Andrei Shleifer, and Robert Vishney, Privatizing Russia (Cambridge: The MIT Press, 1995).

(42) A 1994 survey of some 400 Russian enterprises provides support for these conclusions. For example, privatized firms did indeed have fewer links with the government than state-owned firms. See Simon Commander, Qimiao Fan, and Mark E. Schaffer, Enterprise Restructuring and Economic Policy in Russia (Washington, DC: World Bank, 1996).

(43) These conclusions are based in part on Commander, Fan and Schaffer.

(44) Barbaris et. al.

(45) Georges Korsun and Peter Murrell.

(46) A. H. Shaikh and M. Minovi, Management Contracts: A Review of International Experience, (Washington, DC: World Bank, 1994).

(47) ibid.

(48) Christine Kessides, Institutional Arrangements for the Provision of Infrastructure: A Frame-work for Analysis and Decision-Making, (Washington, DC: World Bank, 1993).

(49) Elliot Berg, "Privatization in sub-Saharan Africa: Results, Prospects, and New Approaches," JoAnn Paulson, ed., The Role of the State in Key Markets, (London: MacMillan, 1996).

(50) Between 1987 and 1995, the percentage of local governments contracting out rose as follows: janitorial services, 52 to 70 percent; garbage collection, 30 to 50 percent; building maintenance, 32 to 42 percent; security services, 27 to 40 percent; street maintenance and repair, 19 to 37 percent and data processing, 16 to 31 percent. (Data from Mercer Group, Inc.).

(51) The generality of the trade-off between speed and quality is indicated by its frequency in the case studies detailed in Christopher Adams, W.P. Cavendish, and Percy Mistry, Adjusting Privatization: Case Studies form Developing Countries, (London: James Currey, 1992).