خصوصی سازی و ساختار بازار در اقتصاد گذار

A model is developed in which an industry of N≥1 firms is privatised. The ‘participation’ method of privatisation is used, whereby firms are sold for cash, but the state retains a proportionate share of ownership. In each firm the new private owner has the opportunity to make a reorganisational investment, before output is produced. This investment is unobservable by the state, and therefore non-contractible. There is Cournot competition in the product market. The welfare-maximising retained ownership share for the state is analysed, taking into account that potential buyers of firms may have limited access to finance.

In the transition economies of Central and Eastern Europe and the former Soviet Union privatisation has taken place by a variety of methods (see, for example, Estrin, 1994 and Brada, 1996). Following the lead of the Czech Republic and Russia, the most common method (used or proposed) has been voucher, or mass, privatisation. However, Germany, Hungary and Estonia have relied heavily on privatisation by sale, with foreign investors playing a major role in the latter two countries. Moreover, from 1995 onwards the Czech and Slovak Republics, Russia and several other countries switched emphasis to privatisation by sale (EBRD, 1996b). In many of these privatisations the state has kept a significant share in the ownership of firms, while giving up all managerial control (see Perotti, 1995, on the Hungarian case). This ‘participation’ model of sale, which is the subject of the present paper, has been forcefully advocated by Sinn and Sinn (1993) and Bolton and Roland (1992). It has also been analysed formally by Demougin and Sinn (1994), on whose work we build. In this model, privatisation is undertaken with two objectives in mind: to bring about investment in the modernisation of firms and to generate revenue for the state.1 The investment is required because of decades of poor technological and organisational achievement under Communism, while state revenue is a critical factor largely because profit tax revenue has collapsed at the same time as some spending needs (for example, for the provision of a social safety net) have risen substantially.2 Unlike Demougin and Sinn, who focus on risk-sharing, we do not allow for uncertainty. This simplification allows us to introduce several other considerations into the analysis. First, in previous theoretical work it does not seem to have been taken into account that the firms being sold may compete against one another in the product market. Yet, both the amount that a buyer is willing to pay for a firm and the willingness of the buyer then to invest in the reorganisation of the firm will depend on how competitive the product market is. In our model this is recognized by supposing that an industry of N firms is being privatised, where N≥1. The buyer of each firm is assumed to make an investment in its reorganisation before production takes place, after which firms play a Cournot production game. Second, we assume that investment by the new private owner of a firm involves the allocation of resources in a way that may be unobservable and non-contractible. In contrast, Demougin and Sinn assume for most of their analysis that there is contractibility, with the amount of investment the buyer will make specified in the contract when the firm is bought. This corresponds to the investment targets that were set for buyers of German SOEs and which have more recently been specified in Estonia and the Slovak Republic (EBRD, 1996b). However, as Demougin and Sinn point out, a government will be unable to observe the real cost to a company of transferring managerial know-how to an acquired firm, for this depends on the managers’ alternative occupations. In our model the amount of investment is chosen freely by the buyer and may not be observable to outsiders. Hence, we suppose that the full cost of investment is borne by the buyer.3 The government’s participation in the firm relates only to the share it takes of production profit (the government is a sleeping partner). This participation may be interpreted as ownership or as a cash-flow tax. A cash flow-tax may be less open to abuse than a profit tax would be.4 Third, we investigate two different forms of reorganisational investment. One form updates methods, reducing marginal cost for a good that is already in production. The other form creates capacity to produce a new output, for example, as Volkswagen has created capacity in Skoda to produce cars of, for Skoda, a previously unachieved quality. For the latter form of investment we also allow for the possibility that the good is internationally traded, with producers facing a horizontal demand curve. Fourth, we allow for the possibility that any potential buyer of a firm may be financially constrained. This is to reflect the fact that in transition economies the main source of investment funds, domestic savings, has declined sharply in real terms, while the fragility of the banking sector undermines savings mobilisation and financial intermediation. Although there has been a recent increase in sales of firms to foreign companies, this has been concentrated in a few of the transition economies and in particular market segments (EBRD, 1996b). Furthermore, foreign companies also have limits on the funds they have available.5 In our model the finance constraint plays two potential roles. It may prevent a buyer from paying the amount the firm is worth; and given the amount it pays for the firm, the buyer may have insufficient access to funds to raise the amount of reorganisation investment to the profit-maximising level. We begin our analysis by formulating a ‘basic model’ in which there is no binding constraint on finance. Given that the industry faces a downward-sloping demand curve, then, if government revenue is given a large enough weight in the welfare function, the optimum retained ownership share for the state tends to be larger when there are more firms in the industry being privatised. This conclusion is unaffected by whether reorganisational investment occurs as cost reduction or capacity expansion. However, for the latter case, we also allow for the effect of international trade through the alternative assumption that the goods demand curve is horizontal. Then, we find that, provided the industry is commercially viable, the state should not keep any share in ownership. When we introduce a binding limit on finance into the model, we find that, generally speaking, the more finance is limited, the greater is the share in ownership that the state should keep. Also, for the case of investment in capacity expansion we describe a possible situation in which a limit on finance leads to the non-existence of an equilibrium price for a firm. Since we assume throughout that any retained ownership in a firm by the state is in the form of non-voting shares, the government cannot control a firm’s behaviour directly. Nonetheless, the government can influence a firm’s behaviour through its choice of how big an ownership share to keep. The underlying feature of the model that generates our main results is the tendency, when finance is unlimited, of Cournot competitors to overinvest relative to the collusive profit-maximising solution. When, however, the government takes a share of production profit, the incentive to invest is less for the private owner, ceteris paribus. This can raise production profit for the industry and revenue for the government (from its ownership share, if any, plus the price it is able to get for the share it sells). The government should take an ownership share if it values revenue sufficiently greatly, compared with investment (or consumer surplus). If a limit on finance for firms prevents such a solution from being attained, the optimal government ownership share tends to be yet higher because this causes the price that a private buyer is willing to pay for its share of a firm to be lower. The private buyer is therefore left with a larger proportion of its financial resources to use for investment, limiting the fall of investment below the level in the finance-unconstrained solution. If there were no binding constraints on the availability of finance or on the value that the private ownership share could take, a first-best solution would always be achieved. However, when such constraints bind we enter a second-best world. A constraint that we impose throughout is that the private ownership share cannot exceed unity, i.e. the state will not subsidise production profit. This constraint binds when the government values investment greatly compared to revenue. Another potentially binding constraint, the effects of which we note in the final section, is that for ‘privatisation’ to take place the private ownership share may have to be at least one-half. Before proceeding, it is worth considering the possible empirical relevance of the potential for overinvestment when oligopolies are privatised, though we can only draw indirect inferences. First, note that, according to EBRD (1995), the ratio of gross fixed investment to GDP for 1993–1994 was ‘fairly high’ in many transition economies, compared to the OECD average (for 1994) of 20.6%. For example, for the Czech and Slovak Republics, Slovenia, Hungary, Estonia, Lithuania, Belarus and Russia the ratio was in the range 20–28%. And EBRD (1996a) reports that the ratio of investment to GDP tended to rise from 1994 onwards outside the CIS.6 It should also be emphasized that much investment is in forms that are not recorded in official statistics. For example, as we have already mentioned, there is an opportunity cost to transferring managerial resources from the other activities of a company.7 This may be especially important for foreign investors. Second, we may focus on Hungary and Estonia, since, apart from the special case of Germany, these are the main exponents of privatisation by direct sale. Duponcel (1998) reports that in the Hungarian and Estonian food sectors foreign direct investment has considerably increased competition, while EBRD (1997) notes that in Estonia the contractual investment obligations specified by the government when selling firms have often been exceeded, sometimes by a wide margin. Though such evidence is only circumstantial, it is consistent with the hypothesis that investment in some industries may be greater than the amount that would maximise government revenue. We begin our analysis by examining the ‘basic’ case, in which finance is unconstrained. The model is set up for this case in Section 2 and solved in Section 3. Then, in Section 4, we introduce a shortage of finance. Section 5 concludes by discussing the implications of changing our set of assumptions. Appendix A provides proofs of propositions and deals with the technical points.