Integration, internalisation and market failure

In this section, we consider integration in the context of market failure. The analysis runs parallel to that of Chapter 6 on the policing of labour con­tracts, but here the emphasis is on the problem of contracting with buyers or suppliers of non-labour inputs. Integration as a means of overcoming transactional hazards is sometimes called the ‘internalisation’ approach to the firm. Where markets fail, an alternative form of organisation is substi­tuted and the market transaction is ‘internalised’ within the firm. Usually, this process is seen as increasing economic efficiency. Subsection 5.2, however, outlines cases in which the pursuit of monopoly advantages can be viewed in a transactions costs framework and may or may not lead to greater economic efficiency.

1. Monitoring Input Quality

If the quality of inputs is observable at very low cost, we could expect market transactions to be used for recurrent purchases of ‘standard’ items.18 Even where quality is costly to monitor, full integration within the firm is only one of several options available. The continuing use of a par­ticular supplier with a good ‘reputation’ to protect is a possible response to adverse selection. ‘Cheating’ on the part of such a supplier would be costly if discovered and ‘the discipline of continuous dealings’ with the implied threat of terminating the association following unsatisfactory performance may provide a ‘solution’ to the incentive problem. Clearly, this relatively simple response to adverse selection and moral hazard requires both the existence of ‘reputable’ suppliers, and at least some ‘informative signal’ about quality in order that a ‘monitoring gamble’ can be constructed. Obviously, the more informative the signal used, the more effective the monitoring of the input will be. Information is valuable to the buyer and may be sought in a variety of ways depending upon the costs involved.

It may be, for example, that simple inspection of the input is sufficient to ascertain its quality and reliability, and that staff with specialist knowledge can be employed simply to undertake this task. On the other hand, inspec­tion of the input itself may not provide, at reasonable cost, information about quality. Knowledge of production conditions may be important. If this is so, a number of consequences follow. The first possibility is that detailed environmental circumstances, to which all producers are subject, may have important effects, and that these can be appreciated only by prac­tical experience. This might result in a buyer integrating backwards into the production of an input simply to acquire this kind of information and thus to remove the information asymmetry associated with dealings with other suppliers. The objective is not to produce one hundred per cent of require­ments of the input, but to produce information. An example already quoted is that of the interest of Guinness in malt production. Similarly, the history of Sainsbury provides a number of cases: ‘Before the war we were very large buyers of beef cattle in Aberdeenshire, through our Northern agent, and it was our desire to get direct experience of rearing and fatten­ing costs . . . that led to our farming enterprise.’19 A second possibility is that quality is determined primarily by the efficiency of the production operation run by the supplier. This might lead to arrangements for techni­cal inspection and direct monitoring between independent firms. Finally, where the detailed specification of the input is continually subject to change, technical inspection may give way to full internal integration within a single firm.

Forward vertical integration of manufacturers into retailing involves a similar range of options. Where one manufacturer’s product is indistin­guishable from another’s, the use of market contracts and an independent wholesaler is suggested. The wholesaler develops contacts with retail outlets and supplies them with a variety of products from his stock. The transactional economies involved in employing an intermediary are clear since the alternative is that each manufacturer must develop links with each retailer. Problems arise, however, when it is no longer a matter of indifference to the manufacturer where his goods are sold, how they are treated and how presented. Faulty handling or poor presentation in the wrong environment may affect the brand image of the manufacturer’s product. Complex products or new products require that consumer confidence is established, and once again this involves the development of a ‘reputation’ which is attached to a brand name. Manufacturers may respond to this problem in a number of ways. The first option is to estab­lish a sales force to sell direct to retailers and monitor performance, pre­sentation and general quality of the store. This is the policy adopted by Yardley and Co., the manufacturers of perfumery and cosmetics in the 1930s and 1940s. Selling more complex commodities may require a knowl­edgeable person to communicate with each potential customer, and, here, forward integration by the manufacturer can range from the direct training of the sales force and the use of space in a department store or (more recently) ‘shopping mall’ through franchising, to full integration.

Etgar (1978) reports that full or ‘partial’ vertical integration (franchising) into retailing is common in the USA. It is a system through which ‘more than one-third of products and services is currently distributed’ (p. 249). He argues that forward vertical integration is ‘motivated by a desire to achieve product differentiation in the ultimate market (emphasis in original). If a distributor represents several suppliers, all will benefit by better service, whereas exclusive distribution is required to tie a good image to a particu­lar brand. Etgar therefore formulates the hypothesis that ‘suppliers who forward integrate from a competitive level into a competitively structured distributive level will provide more services’ (p. 251). This hypothesis is tested using data from the property and casualty insurance industry. In this industry, the system of independent insurance agents has been challenged since the war in the USA by the direct writing system (using either employ­ees or franchisees). Etgar isolates twenty-one service variables (such as degree of home inspection, and time taken to settle the majority of claims) and compares the performance of the vertically integrated and non-inte­grated distributors. Significant differences were detected in eleven of the service areas, and eight out of these eleven service variables showed a higher quality performance by vertically integrated distributors.

2. Internalisation and the Multinational Firm

In our discussion of the economics of the chain store in section 4 we emphasised economies in the use of information gathered throughout the chain. This leads conveniently to the problem of the multinational firm. In both cases, the same fundamental phenomenon requires explanation – the integration of geographically dispersed operations in a single firm. Traditional theory has addressed the problem of explaining the geograph­ical dispersion of production (for example, Heilbrun, 1981). In essence, the analysis derives from the location theory of Weber (1929), Alonso (1964) and others, as well as the ‘central place theory’ associated with Christaller (1933) and Losch (1954). In this theory, the advantages of concentration (mainly economies of scale) are traded off against the advantages of dis­persion (mainly savings in transport costs), and a ‘least cost’ spatial pattern of establishments is derived for a given distribution of demand over the area. There is no clue here, however, as to why the establishments so dis­tributed are integrated within a single firm in some instances but not in others. Similarly, in the international setting, static theory does not explain why production is undertaken by indigenous firms in some cases and by affiliates of a multinational corporation in others.

Dunning (1973) surveys not only location theory but also traditional trade theory in an attempt to highlight the ultimate source of multina­tional enterprise. He concludes that neither have any relevance to the issue. If the affiliates of a foreign firm have advantages over indigenous firms, this must have something to do not with advantages associated with the country concerned but rather with advantages specific to the multinational enterprise. The character of these advantages is important: ‘Essentially, they are enterprise-specific, that is they are not transferable between firms, and are a function of their character and ownership’ (p.314). The multi­national enterprise, according to this view, possesses a resource not avail­able to the indigenous firms and for which there is no effective market, so that acquisition of the resource through the process of exchange is ruled out. Special skills and technical ‘know-how’ clearly come into this cate­gory. Knowledge acquired by pure experience at doing the job is entrepre­neurial in character and impossible to market. It is specific to the firm because it is generated by the entrepreneurial talent of the people who happen to be working in it. Use of this firm-specific information may then result in multinational expansion.

Dunning’s explanation of multinational production is sometimes called the ‘eclectic’ or ‘Ownership-Location-Internalisation’ (OLI) theory. Location advantages are necessary simply because without some reason to locate facilities in differing countries a firm would simply produce at a single location and export to the various markets around the world. Clearly transport costs, tariff or other barriers, economies of scale or varying factor prices give rise to location advantages. Ownership advantages are necessary because without them the firms located around the world could all be inde­pendent of one another. The exploitation of a firm-specific resource, such as a patent, a brand name, a set of marketing or production skills, man­agerial and organisational structures or ‘know-how’, helps to explain why firms located in many different countries are under common ownership. Internalisation advantages are necessary to explain why a firm with own­ership advantages does not simply capitalise on them by contract rather than ownership. In other words the firm might license the use of its patented techniques or set up management advisory services to market its organisational expertise. Some ‘market failure’ or prohibitive transactional hazard explains the use of the firm rather than the market.20 Further dis­cussion of this approach to the multinational can be found in section 6.1.2.

3. Integration and Market Power

3.1. Monitoring restrictive agreements

Integration in pursuit of monopoly profits has been considered a primary determinant of industrial structure for many years. Even in this familiar area, however, it is necessary to understand the implicit assumptions about transactions costs which underlie the conventional theory. The advantages to be gained from ‘combinations’ have been recognised in every age. In a famous passage, Adam Smith writes that ‘people of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices’ (p. 130). ‘Integration’, however, can take many forms other than the creation of a single firm, and the choice of institutional arrangements will depend upon transactions costs. A voluntary agreement between all existing pro­ducers covering prices to be charged and quotas to be produced would fulfil the same objectives as a complete merger. In a world in which information were perfect and transactions costs zero there would be no reason for pre­ferring one method of monopoly creation over another from the point of view of the producers. If integration within a single firm is the strategy fol­lowed, this will be because the costs of monitoring and hence of ensuring compliance in the market exceed the costs of internal organisation.

As Stigler (1964) remarks ‘no conspiracy can neglect the problem of enforcement’ and he bases his theory of oligopoly squarely on the problem of acquiring information about the behaviour of other independent pro­ducers. A rapid and accurate flow of information about the compliance of others with the provisions of a restrictive agreement will render integration by merger less necessary as a means of securing monopoly advantages. It is interesting to note that economists have traditionally been sceptical about the stability of restrictive agreements and have noted the incentive to cheat on their provisions. This scepticism may explain the suspicion with which horizontal mergers are often regarded, internal control being substi­tuted for external agreements.

3.2. The multinational and market power

Multinational integration may be seen as a means of establishing and con­solidating monopolistic positions. An early proponent of this view was Hymer (1976). A firm with some advantage over foreign competitors would not have to expand into foreign markets if it could sell this advantage (perhaps a patented process or product) at a price determined on a com­petitive market. Where overseas firms are few in number, however, the transactions costs involved in agreeing the terms of a licence could be sub­stantial. In order to capitalise upon a special advantage in overseas markets, integration with a foreign firm might then be necessary. Hymer’s analysis at this point can be seen as a precursor of the more modern trans­actions cost school, although the nature of the advantages possessed by a multinational firm would be more likely to be described as ‘competitive’ than ‘monopolistic’ by members of this school. Hymer’s awareness of transactional considerations is emphasised in a recent contribution by Yamin (1991, pp. 74-5).

A second motivation for multinational expansion is simply to remove the potential for conflict with competing suppliers. Again, Hymer emphasised that with very numerous competitors and standardised products multina­tional links between firms would not be so advantageous. With a few oli­gopolists, however, agreements to limit damaging price wars would be sought and collusion across national boundaries could result in higher joint profits. However, enforcing such collusive agreements requires good infor­mation about a rival’s behaviour, and a full merger which creates a multi­national firm can be seen as a method of overcoming this problem. A more recent and extended elaboration of this view can be found in Cowling and Sugden (1987). ‘Transnationals arise because they are a means of consoli­dating or increasing profits in an oligopoly world’ (p. 20).

The possible collusive advantages of multinational expansion are not confined to the product market. Building on the work of Marglin (1974), Cowling and Sugden (1987, pp.62-70) suggest that multinational produc­tion gives a firm bargaining advantages over workers in the labour market. For Marglin, the demise of the ‘putting-out system’ and the rise of the factory system in England at the end of the eighteenth century was not simply the substitution of a more for a less efficient mode of production. Factory production gave the entrepreneur and capitalist much greater ‘control’ over the workforce. No doubt, there were important efficiency consequences, but there were also great distributional consequences. Sugden (1991) argues that the multinational enterprise also has notable dis­tributional effects. In particular, the multinational can threaten workers in one country with the location of new investment in another country. This strategy is termed ‘divide and rule’. ‘By dividing workers into country- specific groups employers improve their bargaining position, thereby gaining at the expense of workers’ (p. 179, emphasis in original). If the multinational form can arise from a distributional squabble, it may be as reasonable to regard it as a vehicle for rent seeking as a means of capitalis­ing on entrepreneurial alertness. As mentioned in Chapter 6, it is often difficult to distinguish rent seeking from entrepreneurship, and the various interpretations of the multinational encountered here provide another example of this general problem.

3.3. Vertical integration and price discrimination

The costs of monitoring restrictive agreements in the market are also reflected in some other ‘conventional’ explanations of internal integration. Consider the argument that an upstream monopolist will integrate verti­cally with a downstream buyer to correct for the tendency of the indepen­dent buyer to substitute other inputs for those supplied by the monopolist if there are technical opportunities for doing so (Vernon and Graham, 1971; Schmalensee, 1973). A vertically integrated concern would, it is argued, increase combined profits by using more of the internally supplied input in the downstream production process. Figure 7.4 illustrates the point using the traditional isoquant-isocost diagram of neoclassical theory. The input of the monopoly supplier is measured along the horizontal axis and ‘other inputs’ on the vertical axis. Given a monopoly price Pm, the down­stream producer will adopt the combination of inputs at a, where the line of constant outlay (slope -Pm/Pn) is tangential to the isoquant labelled q. An integrated producer will be interested not in the monopoly price but in the marginal cost of production of input m labelled Cm. Clearly, the cost of q will be lower at point b than at point a for the integrated producer, and more of input m will be used.

This argument is impeccable from the point of view of neoclassical theory but, once more, it does not address the question of why market alter­natives to vertical integration are not adopted. If substitution against the monopolist is severe, why does the monopolist not adopt a two-part tariff? Under a two-part tariff, the buyer would pay an initial lump sum and then a price per unit of the input purchased. The system is often used for such items as telephone, gas or electricity services, with a ‘quarterly rental’ for the telephone in addition to a price per call. A sufficiently low price per unit would solve the problem of substitution of other inputs by the buyer, and an appropriate ‘standing charge’ would give the monopolist his or her profit. As Figure 7.4 indicates, the resource savings available from moving to point b imply that some two-part tariff arrangement could benefit both monopolist and buyer. Only the bargaining costs associated with imple­menting it would appear to stand in the way.

The main difficulty of the two-part tariff arrangement is in ensuring that the buyer does not purchase the product on behalf of others and then resell it. If this happened on a large scale, the monopolist would miss out on the lump-sum payments which buyers would otherwise have to pay. This, of course, is a problem faced by any scheme of price discrimination. A price of Pm for the first m2 units of the input purchased (Figure 7.4) and a price of Cm for any further purchases would, for example, be an alternative system which would increase the profits of the monopolist seller and not harm the buyer, whose total costs would remain unaltered. Once again, resale from one purchaser to another would completely undermine the monopolists’s position.

Forward integration in this case, therefore, is really a method of policing a system of price discrimination which might otherwise not be possible in the market. Formally, it involves the same reasoning as that used by Perry (1980) to explain the process of forward vertical integration by Alcoa in the years 1888-1930. The standard analysis of price discrimination tells us that if there are several different markets for a product with different price elas­ticities of demand, a profit-maximising monopolist will charge a higher price to the buyers with the relatively inelastic demand. Preventing arbi­trage between the markets is the major problem and it is for this purpose that forward vertical integration may be undertaken. By integrating with the buyer who has the most elastic demand, the monopolist can charge a higher price to other buyers without fear of resale. Perry identifies five major markets for aluminium – as a reducing agent used in the production of iron and steel, and as an input in the manufacture of cooking utensils, electric cable, automobile parts and aircraft. On the basis of the price dis­crimination hypothesis, ‘we expect Alcoa to have been more extensively integrated into those downstream industries with relatively more elastic derived demands for primary aluminium’ (p. 44), and this is the pattern which Alcoa apparently followed. The argument, however, depends heavily on unstated assumptions about the transactions costs of alternatives to integration, and other theorists have interpreted the evidence very differently (for example, Silver, 1984, reported in section 6).

4. Transaction-Specific Investments and Opportunistic Recontracting

4.1. Hold-up and vertical integration

In Chapter 6, section 8, the problem of opportunistic recontracting and ‘hold-up’ was discussed in the context of the supply of labour when the latter acquired highly specific skills. A precisely equivalent problem can arise with the supply of other physical inputs or services to the firm. A sup­plier of a specialised component may have to cooperate closely with the purchaser, and over time gains a great deal of detailed knowledge of the particular ‘idiosyncratic’ problems associated with this contractual rela­tionship. In other words, the supplier gains ‘first-mover advantages’ and a complex bargaining situation arises. As was seen in Chapter 6, the financing of this specialised know-how puts one or other contractor at a disadvan­tage. If the buyer, early on in the life of the contract, purchases the services of the seller on the basis of the seller’s initial or pre-experience efficiency, he or she runs the risk of ‘hold-up’ later. The seller could attempt to recon­tract and extort the difference between the value of his/her services and those of the next-best outsider. Assuming that this threat was lifted by the seller ‘posting a bond’ and accepting lower remuneration in the first pre­experience period, the buyer might then act opportunistically by refusing to repay the bond. Contractors wishing to protect a ‘reputation’ will have reasons not to act opportunistically in this way, as was seen in earlier chap­ters. Nevertheless, the greater the appropriable quasi-rents available (that is, the greater the difference between the value of a particular supplier’s ser­vices and the next-best outsider) the greater is the incentive to act oppor­tunistically and the bigger the contractual problem.

Monteverde and Teece (1982a) use this framework to formulate the fol­lowing hypothesis in the context of the automobile industry: ‘The greater is the application of engineering effort associated with the development of any given automobile component, the higher are the expected appropriable quasi rents and, therefore, the greater is the likelihood of vertical integra­tion of production for that component’ (p.207). They measure ‘engineer­ing effort’ by the cost of development of a component and also take account of ‘the degree to which any given component’s design affects the performance or packaging of other components’ (p.210). Each component is coded according to whether it is produced in-house or supplied by an outside contractor. Monteverde and Teece are then able to show statisti­cally that vertical integration and ‘development effort’ are positively and significantly related for their sample of components.

The existence of ‘appropriable quasi rents’ is not a product only of acquired ‘know-how’. It may also derive from transaction-specific physical investment. In order to supply a particular component to a firm, a producer may have to invest in equipment which is more or less specific to this trans­action. Once more this gives rise to quasi-rents because the value of the equipment in an alternative use may be well below its value in the manu­facture of the component for which it was designed. In the extreme case, the equipment may be totally specific to a given transaction, and the alter­native to using it for this purpose is simply to sell it for scrap. An oppor­tunist buyer, under these conditions, might attempt to recontract with the supplier and seize these quasi-rents. Once the physical investment has been made, the supplier is ‘locked-in’ to a particular buyer and therefore faces the danger that the buyer will adjust the terms of the contract unfavourably.

This argument clearly relies for some of its force on the existence of changing, uncertain conditions. In a static world with contractual obliga­tions closely specified over a long period of time, the possibility of ‘recon­tracting’ is ruled out. Property rights are clearly defined and obligations are spelled out as far into the future as is necessary. No problem of bargaining over quasi-rents then arises. As has frequently been asserted, however, the existence of bounded rationality will usually prevent a comprehensive agreement of this type, and the terms must of necessity leave much that is ‘implicit’. Explicit sanctions imposed by a third party in the event of some contravention of the provisions of a contract are often costly to arrange and require very simple, unambiguous contract terms. Most complex con­tracts rely on trust between the parties and the threat of termination and loss of reputation. In these circumstances, the possibility that a buyer will act opportunistically must be a matter of concern to the supplier.

4.2. The case of General Motors and Fisher Body

Klein, Crawford and Alchian (1978) argue that the avoidance of oppor­tunism deriving from specific investments will lead to vertical integration. They use the historical example of relations between General Motors and the Fisher Body Corporation to illustrate the difficulties of market con­tracts and the tendency towards integration. In 1919, the Fisher Body Corporation agreed to supply General Motors with closed car bodies. To do this, Fisher Body had to undertake highly specific investments in body presses and dies. The contract specified a price for car bodies of cost plus 17.6 per cent (no capital costs were included). In 1926, the two companies merged their operations fully. Over the intervening years, the demand for closed bodies had risen above expectations and General Motors felt the price they were paying was too high in the new conditions. Further, because capital costs were not included in the agreement, Fisher had an incentive to use techniques embodying as little capital as possible. Transport costs, on the other hand, were included and this gave Fisher little incentive to relo­cate their operations near General Motors; a move which the latter tried to encourage.

The case of General Motors and Fisher Body is a complex one and involves considerations other than the mere specificity of the capital invest­ment required. Uncertain, changing conditions will favour full integration with a regular supplier, as was argued in section 4. Focusing on the issue of transaction-specific capital alone, it is possible to argue that opportunistic behaviour can be overcome by quasi-vertical integration, and that full inte­gration therefore is not necessary. Thus, Monteverde and Teece (1982b) suggest that ‘what Klein, Crawford and Alchian have offered is not a theory of vertical integration, but a theory of quasi integration’ (p. 323). If specific investment is required, opportunism can be prevented by making the buyer finance it. The buyer of the input acquires property rights in the equipment necessary for its manufacture and grants certain rights of use to the seller. A relationship similar to that of landlord and tenant described in Chapter 4 is established with respect to property rights in the equipment. The seller of the input supplies the resources which are less specific and can be turned quickly to other uses and are thus less subject to ‘hold-up’.

It is worth noticing that, in the case of physical equipment, financing the investment does not subject the buyer to the risk of ‘hold-up’ as it did in the case of ‘know-how’. A buyer that finances the acquisition of non­patentable special skills and ‘know-how’ may be threatened by a seller who then has first-mover advantages. But a buyer who finances physical equip­ment is not subject to this threat, assuming that property rights can be policed and enforced at relatively low cost. The use of the physical assets can be transferred to another supplier in the event of threats from an exist­ing one: the use of non-patentable ‘know-how’ cannot. This is the essential difference between the two cases and explains why, in principle, transaction- specific physical capital leads to quasi-vertical integration and not neces­sarily to full vertical integration.

The GM and Fisher Body merger has become one of the most inten­sively scrutinised ‘case studies’ in organisational economics. Some writers, for example Coase (2000), deny that Fisher Body ever made an attempt at ‘hold-up’ before 1926. Casadesus-Masanell and Spulber (2000) argue that coordination of production and inventories was a more important consid­eration in motivating the final merger (the forces discussed in section 4 above) and that transactional relations between 1919 and 1926 were char­acterised by a high degree of trust. Freeland (2000) emphasises the impor­tance of access to the human assets represented by the Fisher brothers and points to the transactional problems encountered there. In 1934, the broth­ers threatened to leave if they were not granted options on GM common stock – a successful hold-up of GM which ‘derived from their human assets – their specialized knowledge of the body business and its management’ (p. 58). Against these criticisms, Klein (2000) defends the original interpreta­tion. Relations between Fisher and GM might have been harmonious until 1924 but a sudden increase in vehicle production by 42 per cent in 1925 and another 48 per cent in 1926 tested the original cost-plus agreement to destruction. Fisher’s refusal to develop a body plant at Flint, close to GM’s Buick assembly plant, and its insistence on continuing to supply from Detroit was, for Klein, a clear case of hold-up.

In order to test the theory that specific physical capital encourages quasi­vertical integration, Monteverde and Teece (1982b) took a sample of com­ponents from two divisions of a US supplier of automotive products. Potential quasi-rents were calculated by taking two measures – the simple dollar cost of specialised equipment and an ‘estimate of the percentage of original tooling costs which would be required to convert the tooling to its next best use’ (p.325). They found a significant positive relationship between their measure of appropriable quasi-rents and the occurrence of quasi-vertical integration, but the general explanatory power of their esti­mated equation was low.

5. Enforcing intertemporal commitments

Subsection 5.4 was concerned with the problem that the parties to a con­tract might attempt to ‘renegotiate’ it as time passed. A particularly extreme example would be that of a buyer who terminates the association and contracts with a new supplier after a certain period of time has elapsed. Considerable attention has been given to the incentive properties of termi­nating an agreement in earlier sections and chapters, and it may therefore come as a surprise that there are circumstances in which a buyer may wish voluntarily to ‘disarm’ and forgo this threat of termination. It may, in other words, be necessary for the buyer to ‘tie him or herself down’ and promise not to use alternative suppliers. Clearly, a buyer who does this will wish to monitor the operations of his or her supplier very closely and vertical inte­gration is the likely result.

Goldberg (1976a) argues that granting a producer the ‘right to serve’ a constituency may be in the interests of the buyers if uncertainty about the introduction of new technology and the possible obsolescence of specific capital equipment renders the producers unwilling to undertake the required investment in the absence of some assurance about continuing outlets. The problem is not that the buyer might attempt to capture the quasi-rents associated with transaction-specific capital (as under subsection 5.4), but that s/he might desert the supplier entirely in favour of a new entrant using superior technology. Although, in the short run, it will always be in the buyer’s interests to maintain the freedom to contract with anyone who offers the most favourable terms, ‘the effective achievement of their long-term interests requires that barriers be erected to their pursuit of short- run self-interest’ (Goldberg, 1976a, p.433). Unfettered freedom to contract with anyone else at any time will not be in the buyer’s interests if the sup­plier is thereby rendered unwilling to undertake the necessary investment.21

These ideas were developed primarily as an approach to the theory of regulation and were elaborated more formally by Ekelund and Higgins (1982). Public regulation, particularly of ‘natural monopoly’ markets, sometimes involves restrictions on new entrants, and this has been widely criticised by economists as suppressing competition and innovation. Goldberg does not dispute that giving the supplier a ‘right to serve’ could adversely affect incentives and suppress innovation, but in a very dynamic uncertain environment the total absence of such a right could also suppress innovation and prevent entry. He points out that private contractors will often voluntarily attempt to restrict their future options through long-term arrangements. Interpreted in this way, the firm operates like a ‘thermal’ type of nuclear reactor, increasing the chance that innovations will be ‘captured’ by using a ‘moderator’ to slow them down. Innovations which are, so to speak, ‘uncontrolled’ may be lost. It is a paradox that Schumpeter’s ‘gale of creative destruction’ if too severe, could lay waste all plans to innovate.22 The firm encourages change by simultaneously moderating its force, and assists new entry by restricting the scope for further entry. Further discus­sion of this question in the context of government regulation can be found in Chapter 15.

That restrictions on future freedom of contract do not necessarily prevent entry can be seen by considering the problem faced by a firm wishing to encourage the production of a component at present being sup­plied by a monopolist. It could be argued that the downstream purchaser will have an incentive to produce the upstream component within the firm in order to obtain it more cheaply. The problem with such reasoning is that it is not clear why if the downstream firm is capable of producing the com­ponent in this way, some other independent supplier cannot enter the market and compete with the monopolist. One possible answer is that the new entrant will take time to learn the techniques of production involved, and the established monopolist may have firm-specific cost advantages (that is, there are barriers to entry). In this situation, faced with the monop­olist’s threat to retaliate by cutting price in the event of new competitors entering the market, a potential new supplier would look for assurances from the downstream buyer that a continuing outlet would be provided at agreed prices. ‘Firm-like’ long-term arrangements therefore emerge, and possibly full vertical integration.

Source: Ricketts Martin (2002), The Economics of Business Enterprise: An Introduction to Economic Organisation and the Theory of the Firm, Edward Elgar Pub; 3rd edition.

One thought on “Integration, internalisation and market failure

Leave a Reply

Your email address will not be published.