«Journal of Public Economics 88 (2003) 305 – 332 Privatisation around the world: evidence from panel data Bernardo ...»
Journal of Public Economics 88 (2003) 305 – 332
Privatisation around the world: evidence from
Bernardo Bortolotti a, Marcella Fantini b, Domenico Siniscalco c,*
Universita di Torino and Fondazione Eni Enrico Mattei, 20123 Milan, Italy
National Economic Research Associates, Rome, Italy
The Italian Treasury, CESifo, and Fondazione Eni Enrico Mattei, Via XX Settembre 97, 00187 Rome, Italy
Received 1 April 1998; received in revised form 30 July 2002; accepted 30 July 2002 Abstract Why do countries privatise? This paper presents new evidence from a panel of 34 countries over the 1977 – 1999 period. The empirical analysis shows that privatisation takes place typically in wealthy democracies, encumbered by high public debt, but endowed with deep and liquid stock markets. Budget and ‘market’ constraints matter, but legal institutions are also important. Indeed, the extent of privatisation in terms of revenues and stakes sold appears more limited in civil law countries, where shareholders are poorly protected, banks powerful, and capital markets less developed.
D 2002 Elsevier B.V. All rights reserved.
JEL classification: L33; D72; G15; H6; K22 Keywords: Privatisation; Public finance; Political economy; Law and finance; Capital markets
1. Introduction Privatisation, defined as the transfer of ownership rights of State-owned enterprise (SOE) to the private sector, is a major trend all over the world. The process began in the late 1970s, with the Thatcher government in Great Britain, and spread across countries and ` continents to become a distinguishing feature of fin de siecle capitalism. Privatisations are now common to most countries and occur across geographical regions and sectors. From * Corresponding author. Tel.: +39-06-4761-4189; fax: +39-06-4873-414.
E-mail address: email@example.com (D. Siniscalco).
0047-2727/$ - see front matter D 2002 Elsevier B.V. All rights reserved.
doi:10.1016/S0047-2727(02)00161-5 306 B. Bortolotti et al. / Journal of Public Economics 88 (2003) 305–332 1977 to 1999, 2459 deals in 121 countries worth approximately US$1110 bn were reported. Global SOE value added decreased on average from 9 to 6% of GDP in the 1978– 1991 period (World Bank, 1995). Privatisation also had a tremendous impact on financial markets: by the middle of 2000 privatised SOEs boasted a market capitalisation worth US$3.31trn (Megginson and Netter, 2001).
The empirical literature has provided systematic evidence that privately-owned companies outperform SOEs, and that privatisation enhances the financial and operating performance of firms (Dewenter and Malatesta, 2001; D’Souza and Megginson, 2000).
Despite the large welfare gains that could stem from privatisation, few governments have completely transferred ownership and control of SOEs to the private sector. In the reported public offerings between 1977 and 1999, the majority of stock was sold in only 30% of the 617 companies being considered, and it never happened in 11 out of 76 countries. This rough evidence indicates that control is still very much in State hands and that partial or incomplete sales are a common feature of privatisation processes.
Why do governments privatise? Why do some countries accomplish large scale privatisation programmes, and others never privatise at all? Moreover, how do governments privatise? Why do some governments privatise big stakes in SOEs, while others stick to partial privatisation?
This paper provides some answers to these important questions, implementing a twostage empirical analysis on a panel of 34 developed and less developed economies over the 1977– 1999 period. At the first stage, we try to explain why some governments privatise, and others do not. At the second stage, we estimate the extent of privatisation in terms of the economic value of the assets transferred to the private sector, and of the percentages of capital sold in SOEs.
Our main results can be summarised as follows. The first stage of the empirical analysis shows that, as theory predicts, privatisation is associated with high levels of public debt, a well-functioning domestic stock market, and a right-wing majority in office. First, fiscal imbalances trigger privatisation, as the windfall revenue can be used to square public finances. Second, incumbent governments take advantage of hot markets to float SOEs.
Indeed, a liquid stock market allows divesting governments to obtain the full market value of the company sold, and to generate more revenue from the sales. Third, right-wing governments resort to privatisation in order to diffuse ‘popular capitalism’, achieving the political objective of increasing the support for market oriented platforms.
The first stage identifies possible reasons why some countries do not privatise. Less established democracies with weak political systems appear barely able to set SOE divestiture in motion. The soundness of political institutions is a key component of sovereign risk, which in turn is a priced factor. Therefore privatisation becomes less feasible in less democratic settings, as governments are forced to implement highly discounted fixed-price offerings. Furthermore, privatisation seems less likely to occur in German civil law countries, such as Austria, Germany, Japan, South Korea, Switzerland, and Taiwan. Interestingly, all these countries have bank-dominated financial systems.
Banks may have a vested interest in financing SOE with soft budget constraints, and possibly may obstruct privatisation to preserve the status quo.
Privatising and non-privatising countries emerge as two sharply distinct groups, whose differences hinge upon the economic, political, and institutional environments where B. Bortolotti et al. / Journal of Public Economics 88 (2003) 305–332 307 governments operate. But once the privatisation decision is taken, why does the extent of privatisation vary so much across countries?
The second stage of the empirical analysis shows the value of the shares privatised relative to GDP—our first proxy for the size of one country’s privatisation—to be affected by domestic stock market development. A deep and liquid market allows the absorption of big issues, so that larger SOEs (and larger chunks of capital of these SOEs) can be more easily privatised. Furthermore, by producing information, market liquidity facilitates monitoring, increases the market value of the company, and allows the divesting shareholder to raise more proceeds from the sales.
Clearly, revenues are useful in providing a first measure of the economic impact of privatisation. Nevertheless, by focusing only on revenues one of the key question in privatisation remains unexplained: Did ownership change hands? To address this question, it is only natural to look at the percentage of capital sold to private investors—our second proxy for the extent of one country’s privatisation. Now, legal institutions play a role.
Indeed, the empirical analysis shows that the transfer of ownership (and possibly control) appears more limited, and therefore privatisation more partial, in French civil law countries as opposed to common law countries. The law and finance literature has shown that the French civil law origin is associated with poor minority shareholder protection. Legal protection matters also in the context of privatisation, as government should care about the class of newly created shareholders being expropriated by the managers of privatised SOEs. As a consequence, where the law affords weak protection to shareholders, governments are more reluctant to relinquish control, and privatisation remains partial.
From the 1980s onwards, privatisation has inspired an extensive empirical literature, and has now become an established field of research (see Megginson and Netter, 2001, for a comprehensive survey). However, to our knowledge our paper is the first multi-national study dealing with the determinants of privatisation using panel data analysis over a long period of time. Few empirical papers have dealt with the issue using cross-sectional data.
Bortolotti et al. (2001) provides first evidence that privatisation is affected by the political majority, budget deficits, and legal institutions. Jones et al. (1999) study underpricing in 137 privatised companies in 34 countries and find evidence that it is more frequent where governments need to gain domestic political support. Megginson et al. (2000) study the choice of a private placement versus flotation on public equity markets in 1992 privatisations in 92 countries, finding that the frequency of share offerings is positively related to the size of the firm.
The paper is organised as follows: Section 2 states the theoretical hypotheses being tested; Section 3 describes the data; Section 4 presents the empirical methodology and the results of the econometric analysis. Section 5 concludes.
2. Determinants of privatisation
Which factors explain privatisation across all countries? This section describes the theories that we assess. The possible determinants of privatisation we focus on are classified into four groups: (i) political preferences; (ii) hard budget constraints; (iii) legal origin; (iv) stock market liquidity.
308 B. Bortolotti et al. / Journal of Public Economics 88 (2003) 305–332
2.1. Political preferences
It is often argued that privatisation has a political dimension. Conservative parties are believed to be more prone to privatise the economy than socialist or Christian-democratic parties. Indeed, large scale privatisation programmes have been often associated with the leadership of ‘right-wing’ market-oriented politicians. And the Thatcher’s government in UK is the typical example.
But why should a right-wing government privatise? A rationale for the choice may be a forward-looking behaviour of market oriented politicians aiming at gaining future support from the constituencies of shareholders of newly privatised firms.
Biais and Perotti (2002) formalise this intuition in a bi-partisan model of privatisation where two parties cannot commit to a platform before election. In this context, the rightwing party maximises the utility of the rich, the left the utility of the poor, and each party needs the vote of the median class to win the elections. They show that by allocating a substantial amount of shares of privatised companies to the middle class, the right makes the median voter averse to the redistribution policies of the left, and more prone to vote with the right at future elections. A large scale privatisation program may therefore represent a strategy for switching to forms of ‘popular capitalism’ by creating a constituency of voters interested in the maximisation of the value of their financial assets.
Importantly, as the propensity to buy shares is increasing in wealth, strategic underpricing might be necessary to ensure the participation of the middle class when income inequality is high.
Another important dimension in the ‘political economy’ of privatisation is the government’s credibility, or ability to marshal the support of private investors. This ability is related to many factors, such as reputation of the government, the presence of restraints on policy reversals and on the implementation of economic policies, etc. Credibility is considered crucial for the financial success of privatisation, since it could affect an investor’s willingness to pay (Kikeri et al., 1992). A credible government should therefore be associated with more sales and more privatisation revenue.
Credibility may also affect the size of the stakes privatised. Perotti (1995) provides a theory of partial privatisation based on strategic commitment where, the structure of the offer conveys information on the willingness of governments to bear residual risk. Partial privatisations therefore commit governments not to shift policy in the future. The testable implication of this theory is that a credible government does not need to signal commitment and will be able to sell larger stakes in privatised firms.
Right-wing governments are typically associated with enhanced commitment to market oriented platforms and credibility. Then, the political theories of privatisation yield the
H1. Ceteris paribus, a right-wing government is more likely to privatise, and it should be associated with higher privatisation revenue, and higher percentages of stock sold.
2.2. Budgetary conditions When a government is in financial distress, the pressure to square public finance provides an incentive to speed up privatisation and restructuring (Roland, 2000; La Porta B. Bortolotti et al. / Journal of Public Economics 88 (2003) 305–332 309 et al., 1999). Privatisation, indeed, has been often recommended as a policy of structural adjustment and stabilisation in developed and less developed economies.
Privatisation contributes directly to balance public finances. First, if inefficient Stateowned enterprises are no longer financed by the government after privatisation, subsidies and transfers are cut, with a reduction of expenditures, and an improvement in the primary deficit. Second, privatisation revenues are typically allocated to the reduction of public outstanding debt, generating lower interest payments. Third, public sector debt instruments (such as debt-equity swaps) have been accepted in payment for shares of privatised companies, especially in heavily indebted countries like Mexico, and the Philippines. In this way, foreign debt is directly cancelled. Fourth, privatisation proceeds are sometimes used to finance current expenditure, although this policy does not consider the nonrecurring capital nature of the revenue (Guislain, 1997).
Privatisation could also have an indirect effect on public finance. A sustained privatisation program provides a credible signal of policy change, which contributes to reduction of political risk over time (Perotti and Van Oijen, 2001). Indeed, enhanced credibility improves the credit rating for government bonds, generating lower interest payments, and an easier access to capital markets to finance budget deficits.1 A government in financial distress has more incentives to sell. In this context, we should also observe more revenues since a financially distressed government will first sell
more profitable companies. We can therefore state the following empirical implication: