بازار و ارائه عمومی در حضور آثار جانبی سرمایه انسانی

For many goods and services, such as health, education, legal services, police protection, the cost incurred by an individual supplier for providing quality is affected by the human capital of her colleagues. The paper shows that this human capital externality is crucial to determine whether such goods and services should be privately or publicly provided. Public and private provisions give individuals different incentives to acquire human capital, and the paper shows that either may be socially preferable, depending on the nature of the human capital externality: private provision of the final goods and services gives stronger incentives to human capital acquisition (and may therefore be socially preferable) if own human capital and one's colleagues' human capital are substitutes, and if suppliers with high human capital benefit more than suppliers with low human capital from their colleagues' human capital, but not excessively so.

The boundary between public and private provision of goods and services has moved radically in recent decades, which have witnessed a substantial decline of state provision, widespread privatisation programmes, and increasing private involvement in areas traditionally supplied primarily by the public sector, such as health, education, defence, policing and prison services. Shleifer (1998) surveys some of the literature which asks the normative question of which goods and services are best provided by the public sector and which by the private sector. In this vein, we aim here to establish a rigorous theoretical foundation to this question for services1 with the following characteristics. • The quality of the service is important, and human capital is an important component of quality. •While users acquire the service from an organisation, the service itself is supplied by a specific and identifiable individual or a very small team. •There are human capital externalities in the provision of the service: ceteris paribus, an individual supplier's cost of supplying the service depends on the characteristics of her colleagues. Education, health, legal services, the administration of justice, national defence, policing, criminal investigations, prison services, most of music production, civil aviation and many others satisfy these characteristics. Quality is obviously very important for them, and it is heavily influenced by investment in human capital, which often requires years of full time training. The second property is also obvious: patients go to a hospital, but they are treated by a given physician; students attend schools and universities, but each class is taught by an individual teacher; when our house is burgled or our car stolen we call the police, and a specific officer carries out the investigation; we book a flight with an airline, and our plane is flown by a given crew; for legal advice we contact a law partnership, usually employing many lawyers, and our case is assigned to one individual or at most a small team; justice is administered by a large department, but each case is adjudicated by a small court, and so on: this is clearly fundamentally different from goods such as cars for which it is nigh impossible to attribute output to specific individuals. Finally, doctors working in a hospital, teachers in a school, soldiers in a regiment, officers in a police department, and of course professors in a university are all aware of the third bullet point above: the working environment matters enormously for the costs (monetary and non-monetary and net of benefits) of having a job, and the personal characteristics of one's colleagues affect overwhelmingly the working environment.2 We assume that the quality of the service is observable and enforceable. Thus our paper is not in the recent strand of literature which views contract incompleteness as an important determinant of the costs and benefits of private and public provision (Hart et al., 1997). We aim to convince the reader that human capital externalities are a very plausible explanation, both in theory and in practice, for differences in private and public provision, alongside, and fully compatibly with, contract incompleteness. In the model, the government chooses first how to intervene in the market, whether via direct public provision, or indirectly, taxing and subsidising the private market. Because our paper focuses on the technology of production, arguing that it is intrinsically different according to whether the service is supplied by profit maximising units or by units which are directly controlled by a government agency, we define public provision differently from the established literature, where public provision of a service means that it is funded by the public purse, though it may be provided by private agents (see for example Blomquist and Christiansen, 1999). In other words, in this paper, the crucial separation is between finance and provision: goods and services can be publicly financed and privately provided, as is the case in practice for many capital goods such as roads, schools, hospitals, barracks, weapons systems, social housing, and for services such as health and education which many users receive from private providers while their cost is borne by the public purse. Once they know the mode of provision, individual suppliers invest in human capital: they go to university, medical or law school, teacher training college and so on. They are subsequently employed by an organisation, where they supply their service, which is variable in quality. A supplier's cost of supplying quality depends on her own human capital and on the average human capital in her organisation. Our main result is that either private or public provision may be preferred by the government, depending on the nature of the human capital externality. The crucial difference between the two modes of provision is that they determine different incentives for the potential suppliers to acquire human capital, and therefore the aggregate human capital is different with public and private provision. Specifically, if one's own and one's colleagues' human capital are substitutes, and if suppliers with high human capital benefit a little more than suppliers with low human capital from their colleagues' human capital, then private provision gives stronger incentives to human capital acquisition (see Fig. 2), and may therefore be socially preferable. That intervention in private sector provision through taxes and subsidies could be strictly preferred by the government to public provision might appear surprising: one would think that the government should always at least be able to duplicate what the private sector would do. This is not so, however. There is a fundamental difference between public and private provision: when the government provides the service itself, it can alter the structure of the supply, that is the combination of suppliers with high and low human capital it employs in the units it runs. And this creates a trade-off. On the one hand, it constitutes an additional instrument: in private organisations, the combination of suppliers with different human capital is dictated by profit maximisation. On the other hand, crucially, it also determines a disadvantage. The government cannot commit in advance not to use this instrument to pursue its redistributive goals: therefore, before they train to acquire human capital, suppliers can calculate the government's preferred mix in the structure of the supply, and use this calculation to predict the average human capital of their future colleagues in the public sector: this determines their future cost of supply and therefore their current incentive to acquire human capital. As we show, the structure of the supply with direct government provision may provide insufficient incentives to acquire human capital, and the additional instruments that public provision affords the government may not compensate for the weaker incentives to human capital acquisition. Possibly counterintuitively, we find private provision to be more likely to be preferred when the government has a stronger preference for redistribution. This is however a natural consequence of the trade-off highlighted above: in principle, government provision is more efficient, since the best combination of suppliers in the units can be achieved, and quality can be provided at lower cost, but private provision, which guarantees the highest quality to high income users, raises more taxes, which can then be redistributed to lower income consumers. The paper is organised as follows. Section 2 presents the model: its five subsections describe the players–consumers, units, suppliers, and the government–, and the game; in Section 3 we study the subgame with fixed human capital. The incentive to acquire human capital is studied in Section 4. The example in Section 5 illustrates that the government may prefer either mode of provision, depending on the parameters of the model. Section 6 is a brief conclusion.

Broad stylised facts indicate that the mode of provision varies greatly among services with prima facie similar characteristics. Policing and the administration of justice are publicly provided in practically every country, whereas legal representation (even when it is publicly financed through legal aid), management consulting and auditing, and architectural design are typically privately provided. Conversely, services such as health and education are publicly provided in some countries, almost entirely private in others, and partly publicly partly privately provided in yet other countries.16 The paper does not attempt to offer an explanation to this variety, but rather to provide a theoretical framework to address the fundamental question of which mode of provision, public or private, is socially preferable. Our analysis hinges on the effect of the mode of provision on the incentive to acquire human capital. The source of difference between modes of provision lies in the government inability to make a long term commitment to a given pricing and supply structure design. The macroeconomics literature on economic growth has long recognised externalities in human capital as an important factor, e.g., Romer (1986), Lucas (1988) and Barro (1990)17. This has not been reflected in the microeconomic literature on the mode of provision of goods and services which has concentrated recently on investment embodied in physical, rather than in human, capital, and the different efficiency effects of contract incompleteness under private and public provision (examples are Hart et al., 1997 and Bentz et al., 2002; and Bennett and Iossa, 2006). We find that the preferred mode of provision is “fragile”, in the sense that subtle differences in technology or preferences lead to a different answer to the question of the socially preferred mode of provision. This may call for empirical research investigating quite closely the details of the technology of provision, and in particular, the specific way in which human capital affects the cost of provision for individual suppliers, with specific reference to the relative importance of formal training measured here by one's individual t, and the role of the organisation human capital, T.