REGULATORY BARRIERS TO THE DIFFUSION OF INNOVATION: SOME EVIDENCE FROM BUILDING CODES by Sharon Oster Assistant Professor of Economics and John M. Quigley Associate Professor of Economics Yale University New Haven, Connecticut Previous studies, including most prominently the reports of the Douglas and the Kaiser Commissions, have suggested that outmoded local regulation of residential construction has impeded technical progress in the industry. In this paper, we try to identify the determinants of differences across communities in building regulation. In particular, we use as our dependent variable the permissibility of four particular innovations in a cross section of political jurisdictions in 1970 and try to explain this permissibility using variables measuring attributes of building officials, local firms, labor unions, and housing demand. The data was taken from a special survey of local building departments conducted by Fields and Ventre in 1970. Our results indicate that the education of the chief building official and the level of unionization in the area are the two major factors explaining the probability that a jurisdiction will adopt a construction innovation in its code. Key Words: Building official; building regulation; education; housing demand; innovation; regulatory barriers; residential construction; unionization. PREFACE We are indebted to Charles Field and Francis Ventre for making available to us the raw data from a survey they conducted in 1970 and for several useful conversations during the past year. Ventre's unpublished dissertation [22] was extremely helpful in explaining the technical characteristics of the building code provisions explored in this paper. We also acknowledge the helpful comments of colleagues at the Micro Economic Workshop at Yale University and the research assistance of Gail Trask and Robbe Burnstine. This research has been supported by a grant from the Sloan Foundation. INTRODUCTION It is widely alleged that housebuilding in the United States is a "backward industry" compared with either other sectors of the economy or with residential construction abroad. Precise measures of the "backwardness" of the residential construction industry are, however, ambiguous; little data are available which distinguish between housebuilding and other contract construction activities, such as road building.1 Nevertheless, estimates made in the 1950's and 1960's suggest that by either measure of sectoral performance reduced input requirements for the same - - output (i.e., trends in real costs) or increased output for the same inputs (i.e., trends in input productivity) the construction industry has lagged behind other branches of the economy throughout this century. Grebler, Blank, and Winnick [5], for example, concluded that productivity remained constant from about the turn of the century to the mid-fifties, while Denison found an absolute decline in input productivity from the depression onwards [2]. Similarly, Meyerson, Terrett and Wheaton [9] found that the average consumer was able to purchase more housing services in 1929 than in 1955 despite the real increase in general purchasing power during the period. Finally, crude evidence presented by Nelson, Peck and Kalacheck [12] suggests that the scale of Research and Development (R & D) in the contract construction industry is very small; the ratio of R & D expenditures to value added is three and a half times as large for the economy as a whole as for the construction industry [12, pp. 192-195]. The lagging productivity of the residential construction industry and its low level of R & D expenditures, coupled with the importance of housing in the consumer budget, have been cited as a rationale for public intervention to improve efficiency by the Brookings Institution study [12] in 1967 and by two federal commissions, the 1See Sims [15] for a review of these issues. Douglas Commission [11] in 1967 and the Kaiser Commission [14] in 1968. The muchpublicized experimental program, "Operation Breakthrough," sponsored by HUD in the late 1960's was intended to facilitate rapid efficiency gains. Although the precise "cause" of backwardness in the industry is difficult to identify, there are four peculiar characteristics of residential construction activity which may contribute to its relatively low rate of technical progress. First, effective demand for housing is subject to wide fluctuations, produced in part by the vagaries of the credit market. These demand fluctuations may inhibit the adoption of cost-reducing innovations, especially those which would make the production process more capital-intensive and thus more vulnerable to instability in demand. Further, these cyclical fluctuations may bias the research and development process itself, to discourage the exploration of labor-saving innovations which have been the source of much of the observed productivity increases in other sectors of the economy. Second, the small scale of firms in the construction industry may also reduce the incentives for private research and development. Small scale may be particularly problematic if many of the potential innovations in the industry are in organization, systems design, and in the integration of housing components. Here the minimum efficient scale for R & D activity is presumably rather large, and, more importantly, the returns to R & D not easily capturable by a single firm. Third, the merits of a particular idea or potential innovation in housing may be especially hard to evaluate because the performance of any particular innovation in materials, design or construction method depends upon a complex interaction with other parts of the structure. Since the industry is highly fragmented, it may be especially hard for suppliers to judge the potential of an innovation. This too will inhibit the search for innovation. Finally, the fragmentation of the market is reflected, not merely in a large number of small firms operating in local housing markets, but also in a cumbersome regulatory process which relies upon local political divisions to set standards and to enforce regulations in the materials, design performance and safety characteristics of residential structures. The bewildering variation in local regulations may very well mean that potentially profitable innovations are also illegal in many geographical areas. This reduces the scale at which an innovation can be marketed, its profitability, and may further discourage R & D investment. In particular, the variation in regulation may greatly inhibit research and development by suppliers of building materials and capital equipment (even if the supplying firms are themselves large) since the potential market of a successful innovation is restricted by local building codes. - This paper is concerned with the latter two characteristics in particular, with the operation of local building codes. If local variation in codes is indeed a serious obstacle to technical progress in residential construction, then it is of some interest to consider what factors are responsible for such diversity, as well as what other interests may be served by such local regulation. After a cursory review of the evidence on the relationship between efficiency and regulation in home building, we present empirical evidence suggesting some reasons for the observed variation in local building codes. In particular, we examine four specific cost-saving innovations in residential construction, and we investigate the factors associated with their permissibility in local jurisdictions. THE COSTS OF LOCAL BUILDING CODES The regulation of residential construction is typically delegated by states to local jurisdictions which, in turn, enforce standards and specifications governing the erection and construction of buildings. Clearly the enforcement of minimum standards restricts consumer sovereignty in the consumption of housing services. As with other forms of regulation (e.g., drugs and pharmaceuticals), these codes are rationalized as a means of protecting consumers from the consequences of their own ignorance. It seems reasonable to presume that few housing consumers (or drug consumers) are technically trained to evaluate fully the potential hazards of consuming such complex commodities. This is especially true since many critical components of the dwelling unit - the character of the foundation, the wiring, heating and plumbing systems, etc. are difficult to evaluate in the final product. - In principle, of course, a perfectly informed market could obviate this need for regulation by providing optional insurance policies for purchase by the occupants of dwelling units. There is little reason to expect such a complex insurance market to develop, however, given ignorance on the part of consumers (and potential insurers) about the relative probabilities of infrequent, but nevertheless real, injury from structural hazard. Despite this rationale for some form of regulation, there are indications that the fragmented regulatory process acts as a barrier to improved efficiency in housebuilding. The most direct complaint against the operation of building codes is that the system results in unneeded provisions and restrictions which add significantly to the cost of housing by delaying construction of dwelling units and by preventing the use of the most up-to-date and modern materials. Further, it has been alleged that the procedures for modernizing and amending such codes are slow and laborious, lacking in any objective standards, and are dominated by a very small group of local interests.2 - 2These issues are discussed in some detail in the report of the Douglas Commission [11]. There is conflicting evidence on the magnitude of excess costs attributable to variations in building codes. Several studies have suggested that the direct effect of building codes upon construction costs is small. For example, Sherman Maisel's early study (1950) of the San Francisco housing market concluded that an increase of less than one percent in the costs of newly constructed housing was attributable to "known code inefficiencies." [8, pp. 249-250] Richard Muth's econometric analysis of single detached housing conducted in 1968 suggested that locally modified building codes increased average cost by about two percent (as reported in [16], p. 8). Burns and Mittelbach [1], in their report to the Kaiser Commission, analyzed a survey conducted by House and Home (the leading trade journal) in 1958, and suggested that if the 10 most "wasteful practices" required by building codes were eliminated, the average cost saving for single family housing would be 5 to 7.5 percent. "By assuming the provisions [of building codes] are randomly distributed and by taking account of their varying role in communities, "the authors conclude that "...the estimates represent from 1.5 to 3 percent of the price of an average house." [1, p. 102] In expert Several other analysts have come to different conclusions, however. testimony presented to the Kaiser Commission, Ralph J. Johnson [6] concludes that "...in large urban areas, it may be possible to achieve on the order of a 10 to 15 percent reduction in direct construction costs [or 5 to 8.25 percent of selling price by Johnson's calculations] ...if the constraints of codes and restrictive labor practices are removed and if the industry is allowed to produce as efficiently as it knows how" [6, p. 57]. Survey evidence gathered by the Douglas Commission indicated some real cost reductions achievable by mass production under more uniform building codes [11, p. 262]. The estimates indicated that if 21 "excessive requirements" were eliminated, $1838 would be cut from a typical $12,000 FHA insured house. Again, this 15.3% reduction in construction cost (or roughly 13% in sales price, if one-fifth of selling price is composed of the land component) represents the sum of 21 "excessive code requirements," not all of which are necessarily in effect in any particular jurisdiction. The commission report also notes the problems of one home manufacturer who estimated that, to produce a standard product acceptable to the jurisdictions within his six-state market area, would increase costs by $2492 or almost 21 percent. More important than any increased costs directly attributable to the intrinsic aspects of building codes are the production inefficiencies attributable to their lack of uniformity. As noted by Stockfisch: The absence of such consistency [in building code provisions] |