Skip Navigation Planning & Markets
Subscribe Submission Requirements Editorial Board Archive Links Search Home



Appendix: An update to Baltimore findings

Evidence from Baltimore shows that structural fires per capita have gone up substantially, in spite of the fact that building and fire safety regulation has increased. Therefore, there is no evidence that regulation has achieved its objective to reducing these fires (Cobin 1997, pp. 36-46).

The primary reason that a building code was enacted in Baltimore was to "more effectively prevent the spreading of fires" (p. 35). Nevertheless, despite an overall decline in the city's density, the number of structural fires per capita from 1859 to 1994 increased over twenty-five times. Up until 1994, the increase in structural fires has been over fifteen times since the building code ordinance of 1869 was enacted. Meanwhile, the number of pages in the building code grew from one page to nearly fourteen hundred pages, and the number of inspectors climbed from one to one hundred and seventy-eight since 1869 (pp. 37, 42).

However, data received since my original 1995 research was completed suggest that a dramatic decrease in fires has occurred. The more recent data are problematic because good fire department records were not kept after 1975, which seems a bit curious given that technological improvements should have made record keeping increasingly easier and less costly. During my original investigation, I went to the Baltimore Fire Department and reviewed printouts of data from 1989 through 1994. I was thus able to ascertain the number of structural fires for those years and the number of total fires for 1993 and 1994. Upon returning to the Department in early 1999 to update the data set, I was given a figure for 1994 that was substantially lower than the original number I saw earlier, and the data for years 1995 to 1997 I received seemed quite uncharacteristic (if not questionable). I was informed that all records prior to 1994 had been destroyed, including those which I had seen previously, making a recounting impossible. Adding to the odd event, apparently there was a personnel problem in the records and data entry section at the time of my prior research, and this trouble might have led to my unwittingly recording inaccurate information (12).

Notwithstanding this possibility, one could question the most recent data. First, the data through 1975, which is available in the City Hall archives, is not lost or doubtful. If that data is accurate, then the recent data suggest that the number of fires in Baltimore per capita has been sharply reduced. Indeed, there is 63.1 percent drop in structural fires from 1975 to 1997 3/4 7,144 to 2,637. The previous numbers seemed to follow a growing trend and the latter numbers seem very uncharacteristic -- even suspicious. How could fires in the city have dropped so dramatically? Perhaps they did so, but that fact would seem unlikely or at least very surprising. Second, the recent data have a very large variance from year to year which has not occurred in most periods. Such a decline is possible of course, but then why was the 1994 Figure A1 received in 1999 dramatically different from the Figure A1 received in 1995? If that number had matched the previous number, we might conclude that the number of fires had dropped substantially. Instead, we are left with enigmatic circumstances. Third, unlike the recent data, the previous data received from the Baltimore Fire Department for 1989 to 1994 did seem to fit the trend. But even if those figures are now suspect, clearly the 1975 (and prior) figures are not. The trend alteration casts doubt about the most recent data.

Figure A1: Structural fires in Baltimore per 1,000 population.
Figure A1: Structural fires in Baltimore per 1,000 population.

Figure A2: Percent change in structural fires in Baltimore.
Figure A2: Percent change in structural fires in Baltimore.

Figures A1, A2, A3, A4 and A5 illustrate these data (note that for structural fires, the differentiated portions of Figures A1 and A3 are trend estimates that fill gaps in the time series).

Figure A3: Total fires and structural fires in Baltimore.
Figure A3: Total fires and structural fires in Baltimore.

Figure A4: Inspection force and building code size in Baltimore
Figure A4: Inspection force and building code size in Baltimore

Figure A5: Population density in Baltimore
Figure A5: Population density in Baltimore

By the end of 1993, the likelihood of at least one member of a household being affected by a fire in Baltimore has risen over five times since 1890 (p. 45). Therefore, I concluded after my initial study that "the data do not confirm the hypothesis that building codes and building inspectors alleviate negative externalities caused by fire" (p. 40). In the conclusion (pp. 45-46), I noted:

    Summing up, most scholars seem to think that it is at least possible for regulation to improve safety, if not as a major tenet of their analysis at least as a matter of presumption. Likewise, there seems to be widespread acceptance of economic definitions of risk and tradeoffs. Moreover, there is general agreement that present regulatory systems are more or less inefficient and often ineffective, except that a few go further by saying that regulation might actually be harmful. The key disagreements in the literature arise from what to do about this problem, and what areas should be regulated by the government rather than the market. On the one hand, many scholars contend that regulation can be improved by acknowledging and manipulating public risk perceptions, controlling or adapting to the influence of the media, and by employing better techniques such as cost-benefit analysis. Government should regulate all areas where market failure is present, especially when it is the most cost-effective alternative. On the other hand, some say that public choice and knowledge problems serve to defeat the good intentions of such regulation. While some level of government regulation is necessary, the market should provide most (if not all) regulation. To the extent that government does regulate, the inclusion of techniques like cost-benefit analysis and/or the expected value approach is essential.

    This study complements the theory of risk and safety regulation by extending the analysis of Huber, Holcombe, and Saltzman, as well as Nichols, Zeckhauser, Stigler, and to a lesser extent parts of Viscusi. It provides empirical results of regulatory policy which tend to confirm some important implications in them as well as Huber, Holcombe, and Saltzman's conclusions, viz. that public choice and knowledge problems make regulation ineffective and inefficient, such that safety is actually reduced. In theory, it is possible for inefficient government regulation to be effective, but the costs will be very high. However, in Baltimore, it does not appear that costly fire safety regulation has been either effective or efficient. Moreover, the Baltimore result raises comparative questions about how well market-based alternatives might do in contrast to government regulation. Perhaps market-based institutions would better avoid or ameliorate both knowledge and public choice problems, and provide a more robust paradigm for comprehending risk perceptions. The main conclusion of this study is that successful government regulation of building safety is unlikely.

However, my conclusion would have to be revised considerably if indeed the recent data were found to be accurate. For then we would see a large rise in structural fires per capita through 1975 (or 1993), and an apparent example of regulatory failure, followed by a large reduction in such fires and an apparent regulatory success. At this point, we must await evidence and for further research to be completed by the Baltimore Fire Department into the data discrepancies.

However, the evidence from Santiago is not in doubt. And in my view, neither was the original data in Baltimore. If a data problem in fact exists, it is likely with the current data set.

page 21

IndexContinue

USC Seal


Main Page | Subscribe | Submission Requirements | Editorial Board | Archive | Links

PLANNING & MARKETS
http://www-pam.usc.edu/
ISSN 1548-6036

Copyright 1999-2000
University of Southern California
Los Angeles, California 90089-0626
USA