مدل مدیریت کیفیت درون زا

This paper is concerned with product quality, defined as a kind of durability. Existing models of product quality (in the sense considered here) depend on the idea of signaling, itself driven by an informational asymmetry dictated by “Nature.” The paper proposes an alternative approach, which endogenizes the quality management process. A model is developed that is applicable to the markets for consumer durables and for some intermediate goods. Both competitive and monopolistic markets are considered, and some comparative static results are obtained.

In the management literature, “product quality” is defined extremely broadly. The term has been used to refer to safety, availability, maintainability, reliability, usability, and even price (see, e.g., Besterfield, 1986). Generally speaking, quality is best thought of as a characteristic of the product with the property that all consumers prefer more of it to less, at a given price. Some such characteristics will be known to the consumer before purchase, while others will not. For personal computers, for example, the former type of characteristic might include the size of the processor chip (486, Pentium, Pentium II, Pentium III, etc.) or its speed (300 Mhz, 450 Mhz, 500 Mhz, etc.), while the latter type might include product lifetime, repair costs, etc. Most goods have characteristics of both types, though the second notion of quality raises more interesting questions for firms, consumers, regulators, and for economic theory. It is the notion of quality dealt with in this paper. Firms devote considerable resources to influencing the quality of their products. This influence operates at the level of product design, production process, and post-production quality control. The unity of the quality management process is often stressed in the management literature. The distinction between production and quality control decisions is frequently blurred. For example, a firm may seek to raise its production quality by, in effect, demanding tighter quality control from its components suppliers. Thus a production decision in one firm is inseparable from a post-production quality control decision in another. The model developed in this paper incorporates production, quality control, warranty and pricing decisions into the firm’s overall (expected) profit maximizing behaviour. Product design decisions are not considered. In the model presented here, both firms and consumers will be assumed to be ignorant at the moment of purchase, as to the quality of any given product, though they will be assumed to know the probability distribution of quality. Thus, the model is one of imperfect but symmetric information. It will further be assumed that the firm, though just as ignorant as consumers, is less risk-averse. There thus arises a demand, on the part of consumers, for insurance. This might, for example, be provided in the form of a product warranty offered by the firm, or an insurance policy provided jointly with the product. In the case of intermediate goods, “consumers” may be thought of as firms and “warranties” as compensation clauses built into standard supply contracts. Heal (1977) develops a model, involving warranties, which adopts precisely these informational assumptions. He remarks: “Typically the quality control is sufficiently imperfect that no one [i.e., neither seller nor buyer] will know in advance of [a product’s] use what [its] quality will be, and consequently some form of guarantee will be offered.” [Remarks in brackets added.] In Heal’s model, the firm is assumed to produce a probability distribution of qualities which is simply taken as given. He does not seek to model the process by which the firm attempts to alter that distribution. In this paper this process is modeled. In particular, the firm is able to influence the distribution of quality in its marketed output, both by production (choice of technique) decisions and by quality control decisions. It will also be able to offer a product warranty to the market. A standard problem, often assumed away, in the literature on quality, is that of moral hazard on the part of consumers. If consumers can themselves influence the probability or size of a claim under the warranty, for example by failing to take proper care of the good during consumption, then the economic role of warranties may be reduced. See, for example McKean 1970 and Oi 1973, and Priest (1981). For simplicity, moral hazard will be assumed away in this paper. Warranties, whether voluntary or legally compelled, have an important bearing on quality management decisions because the higher the quality of a firm’s marketed output, the lower the likely warranty costs experienced by the firm. Thus, warranties provide the firm with an incentive to market high-quality products. This connection between warranties and quality management has been apparent to managers for some time. Wright (1980), for example, describes events at General Motors: “I instituted a programme for testing and repairing faulty cars as they came off the assembly line—and the results were phenomenal. It cost about $8 a car, which drove The Fourteenth Floor up the wall. But I figured one way or the other we would end up fixing the defects or paying to have them fixed through recall campaigns or dealer warranty bills … The internal quality control audit revealed a 66% improvement in the quality of a Chevrolet coming off the assembly line between 1969 and 1973 models. And most important, warranty costs of our new cars were down substantially.” The existing economics literature deals with both the notions of quality discussed above. When quality is known to consumers before purchase, the focus of interest is screening. The seller will be ignorant as to the preferences of any individual consumer, although he may be assumed to know the distribution of preferences across the population. His problem then is to provide a price-quality schedule, perhaps along with a warranty arrangement, to the market, with a view to screening consumers, and thus extracting the maximum surplus from them. The firm produces a “product line,” deliberately differentiating his product by quality. An obvious example is personal computers: most manufacturers produce a product line involving different processing speeds, amount of RAM, size of hard disk, etc. Authors who develop screening models of quality include Mussa and Rosen (1978) and Matthews and Moore (1987). Signaling models, by contrast deal with a different asymmetry of information, namely one concerning the product itself. Such models are driven by exogenous “type uncertainty.” That is to say “Nature” dictates a firm’s quality, which is then known to that firm but not to consumers. The firm’s problem then is to signal its quality to consumers using price, warranties and possibly advertising. In a repeat purchase framework, the firm may be able to build up a “reputation” for quality. Authors who develop signaling models of quality include Grossman 1981, Milgrom and Roberts 1982, Milgrom and Roberts 1986, Kreps and Wilson 1982, Klein and Leffler 1981 and Shapiro 1983, and McClure and Spector (1991). This paper is concerned with markets in which consumers do not know the quality of individual products before purchase. Screening models do not deal with this notion of quality. Signaling models, by contrast, do analyze this concept of quality, but they do so in a way that treats quality as exogenous and not affected by the firm’s decisions. Both screening and signaling models place the emphasis of the analysis on an asymmetry of information. The approach of this paper is to allow firms’ choice of technique and quality control decisions to influence product quality under conditions of imperfect but symmetric information. Section II describes a theoretical approach to modeling quality management and section III develops this approach into a tractable model. Section IV describes the demand side of the economy, while section V covers the supply side. Equilibrium and comparative statics are discussed in section VI, and section VII concludes. Mathematical details are relegated to an Appendix (equation numbers prefixed with the letter A refer to equations in the Appendix).

It is frequently the case that the quality of individual products is unknown to consumers before purchase. In this situation, the determination of product quality and of warranty contracts need not be modeled as the consequence of type-uncertainty, treating quality as dictated exogenously by “Nature.” For many industries including cars, consumer durables, and some intermediate products, it is plausible to assume that both seller and buyer know the distribution of qualities in the seller’s output, but not the quality of any individual product. The paper presents a model, based on this assumption, in which the determination of product quality is treated as an aspect of the firm’s (expected) profit-maximizing decision. Product quality is taken explicitly to depend upon factors over which actual businesses have control, namely the firm’s choice of technique and its quality control decisions. Choosing higher quality techniques and undertaking quality control are both costly, but they will still be undertaken because warranty commitments provide the appropriate incentive. There need be no legal compulsion to force firms to offer warranties. If buyers are more risk-averse than sellers, a demand for insurance arises, which will induce sellers to offer warranties voluntarily. In the model presented, firms under monopoly and under competition voluntarily offer a “full compensation” warranty. The paper shows that monopolies will take the same quality management decisions as competitive industries, offer the same average quality of output and offer the same warranty deal (a “full compensation” warranty). They will, however, do this at a higher price. In fact, a monopolist will extract all the surplus, while, under competition, it all goes to consumers. The paper also derives some comparative static results that could form the basis for empirical testing. They allow different industries to be distinguished, and they separate long-run from short-run effects. Both supply side and demand side effects are analyzed. The quality management process is complicated, and the model presented here inevitably simplifies that process. Nonetheless, quality management is central to the determination of the quality of marketed output, and the economic analysis of product quality should therefore include a treatment of this aspect of the firm’s behavior.