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

III. Evidence pertaining to fire safety regulation from northeastern Santiago

Santiago, the capital city of country with a booming economy and many years of robust and sustained economic growth, provides fertile soil for examining how well building safety regulation has reduced the number of structural fires per capita. The relatively recent entrance (and use) of fire insurance products, political tensions, and the low incidence of arson adding to Santiago's appeal. Plus, a good data set on fires (which is not always easy to find) covering nearly all of the 20th century was available -- even surpassing the quality of the data set available in when I studied fires in Baltimore (Cobin 1997, chapter 1).

Objective of the Santiago building code

The first building code in Chile, comprising about two pages (6), was legislated after a disastrous earthquake struck Talca in 1928 (a medium-sized city a few hours south of Santiago). The legislation was designed to establish compulsory construction norms for minimum and maximum building height and the selection of building materials, plus a means to retard the propagation of fires and lessen risks due to earthquakes and "other phenomenon". It was also designed to improve hygiene and health conditions, as well as the exterior appearance of buildings (i.e., reduce visual negative externalities).

Fire protections services in northeastern Santiago

There are twenty-two fire departments today in the greater Santiago metropolitan area. The oldest of these, the Cuerpo de Bomberos de Santiago (CBS), was formed in 1863 in response to a tragic downtown factory fire that left two thousand people dead. Beginning in 1903, various parts of the region began to form their own fire departments, with a number of departments springing up (and breaking off from CBS) in southeastern sections of Santiago during the 1930s. Thus, today, CBS services only northeastern Santiago, including the comunas of Santiago, Recoleta, Renca, Independencia, Estación Central, Providencia, Las Condes, Vitacura, and Lo Barnechea (7). The first five of these comunas represent the central and extended downtown area of the city while the last four represent the northeastern and most affluent sectors. Together they comprise the most important commercial zones in Santiago, about half of its land area, and just over one quarter of its population.

Accordingly, the data in this study reflect fires only for those sections of Santiago serviced by CBS. That is, all of extended Santiago up to 1932, and the northeastern sector thereafter. Plus 3/4 though in decreasing numbers 3/4 various smaller comunas and outlying rural regions continued to be covered by CBS until 1963. The CBS data are the oldest in Santiago and probably the most complete available. This data set begins in 1898 (a major fire in the CBS offices in 1891 destroyed most earlier records) and continues nearly without interruption until 1997, although the fire data are missing from 1914 to 1919 and are very incomplete in 1970, the first year of famed socialist President Allende.

As with all archival research, there had to be some interpretation and interpolation in the CBS data set. For instance, the number of structural fires in the periods 1937-1978 and 1987-1994 are not exact and had to be estimated (using one of two methods) from the total fires count using coeval information. The first method simply entailed categorizing each fire according to my best judgment. The second method was to find the percentage of all calls to CBS (including non fire calls, like rescues) that were in buildings and multiply the number of total fires by that percentage. Overall, I believe this effort was quite successful, such that we can be fairly confident that the estimates produced are reliable, only having small difference in most of the figures. Only in a few years were there significant problems, such as unusually large "unclassified" or "unknown" categories, that made it difficult to create a reliable estimate. Moreover, trend or other estimates were used on a few occasions to fill in data gaps, such as the data gap from 1914-1919.

Common CBS jargon posed some difficulties at first, but these were readily overcome. For instance, all fires are not called "fires" by CBS staff. There are "incendios" and then there are "alarmas" or "llamadas a comandancia". Both of these categories are fires. However, there are far less fires in the incendios category, which consists entirely of major conflagrations, whereas the latter category counts all smaller structural or outdoor fires (8). For the purposes of this study, both fire categories were counted and reclassified according to whether they occurred in some sort of structure or not, since only structural fires are relevant to building code considerations.

One important difference between CBS (and other Chilean fire departments) and those in places like Baltimore are the means of financing and managing their activity. CBS receives considerable support from donations and other means in the private sector, although over half of its revenue comes from federal and municipal transfers. However, there are no government employees at the CBS. Its management is completely private. Another important difference is the structure of the labor force. All CBS firefighters are volunteers. Of further interest, not only do these volunteers not get paid to fight fires, they also have to pay to participate in the CBS by covering their own uniform and miscellaneous expenses (9). In addition, employers sometimes frown on hiring CBS volunteers because of the time off they have to take and because of the risks involved to the life and health of such an employee. Apparently, this overall negative net outlay is not a sufficient deterrent to would-be firefighters who apparently enjoy the "thrill" of fighting fires or receive other utility gains.

Figure 7: Volunteers (bars) and total staff (line) per 10,000 population in northeastern Santiago.
Figure 7: Volunteers (bars) and total staff (line) per 10,000 population in northeastern Santiago.

Figure 7 illustrates the number of volunteer firefighters per 10,000 population during the last century (note that the lighter bars represent trend estimates that fill gaps in the time series). The period 1898-1932 includes all of Santiago but 1933-1998 includes only northeastern Santiago (along with some other areas through 1963). These data are augmented by the total number of paid staff (line) in years when data were available. Note that in the early years, some firefighters were paid 3/4 as seen in the difference between total staff and volunteers from 1898-1913.

In 1998, CBS had an average of 1,985 volunteers on call. It also had a paid administrative staff of 130. Thus, the 1998 total staff of 2,115 corresponds to 17.0 staff members for each 10,000 people in northeastern Santiago and just over one staff member for each structural fire. By way of comparison, in 1898 there were 551 volunteers, total staff of 793, 18.9 staff members per 10,000 people, and 52.9 staff members per structural fire. In 1930, the year after fire regulation began, there were 847 volunteers (number of staff unknown), which represents 10.0 men for every 10,000 persons and 4.9 men per structural fire.

These data indicate that while firefighter participation as a percentage of the overall population has remained in a relatively narrow range over the last century, the efficiency of fire fighting has increased dramatically. It seems that fire fighting services are more productive now than ever, doubtless due at least in part to technological improvements, as evinced in Figure 8. In addition, the largely voluntary or market provision of fire safety in Santiago, might tend to mitigate some public choice and knowledge problems, but not entirely, since over half (55.8 percent in 1994) of the CBS budget is derived from governmental sources

Figure 8: Volunteers per structural fire (bars) and per 10,000 population (line) in northeastern Santiago.
Figure 8: Volunteers per structural fire (bars) and per 10,000 population (line) in northeastern Santiago.

Private fire protection services

Scholars have found strong evidence to support the idea that privatizing fire protection services brings considerable efficiency gains without compromising effectiveness. Ole P. Kristensen found that private production of fire protection services is more efficient, considering the case of Denmark where provision is partly private and partly public. Privatizing these services would also improve incentives to improve safety and lower costs (Kristensen 1983, pp. 1, 8). Roger Ahlbrandt found that the most efficient fire protection services are provided privately (and even more effective in terms of response time). A private company in Scottsdale, Arizona produced high quality fire protection services for 47 percent less cost than its bureaucratic counterparts in neighboring towns. This difference is probably due to the fact that bureaucrats maximize their budgets and minimize production difficulties, action which "may be the antithesis of cost minimization" (Ahlbrandt 1973, pp. 1, 3, 6, 14).

Robert W. Poole studied fire protection services in many places across the United States and found that there has been "an overinvestment in fire suppression and an underinvestment in fire prevention" and that the means of financing public fire departments distorts market incentives and is perhaps unequitable (Poole 1980, pp. 305, 306). In America, 91 percent of all firefighters are volunteers in 24,500 fire departments, which creates substantial cost savings, since wages are 90 percent of the total cost of a paid department (pp. 307, 308). Poole found strong evidence that private fire services are far more efficient than, and just as effective as, their public counterparts (pp. 309, 310, 314-315, 325). "The public good argument can be rhetoric that disguises a large and unnecessary subsidy" (p. 315) while private departments or even contracting out many fire department services to the private sector have proven to provide considerable cost reductions (p. 319).

Population and density

Population in the CBS service area has tripled in the last century, and has increased by fifty percent since 1929 when building safety regulation began (824,124 people in 1929 and 1,235,282 in 1997). Accordingly, but unlike Baltimore which had a decrease in population density from the beginning to the end of the study period, population density has steadily climbed in northeastern Santiago, from 434.5 people per square kilometer in 1930 to 985.6 people per square kilometer in 1997, as illustrated in Figure 9. Hence, while Baltimore's population in 1994 and 1919 were roughly equivalent, and its density is declining, northeastern Santiago's population and density have been increasing steadily.

Figure 9: Population density northeastern Santiago (post-1929).
Figure 9: Population density northeastern Santiago (post-1929).

False or unjustified alarms and intentionally set fires in Santiago

Conversations with fire officials in eastern Santiago indicate that arson is rare in Chile. The fire department in northeastern Santiago has kept a fairly good record of intentional fires during the last century, and very few cases of arson have been reported. Nearly all fires that are classified as "intentional" in Santiago have been in trash cans, in barricades or cars (typically during political protests), or in open fields, rather than in buildings. Figure 10 illustrates the data on false or unjustified alarms and intentional fires per capita for the period 1920-1932 in Santiago and, 1933-1997 in northeastern Santiago (including some other areas through 1963).

Figure 10: False and unjustified alarms (bars) with intentionally set fires (line) 3/4 typically not arson 3/4 in northeastern Santiago.
Figure 10: False and unjustified alarms (bars) with intentionally set fires (line) 3/4 typically not arson 3/4 in northeastern Santiago.

False alarms often pertain to technical equipment failures or mandatory runs to fulfill police mandates. Unjustified alarms occur when a person reports a fire but in reality there is none. It is somewhat impressive that there have been so few false alarms, unjustified alarms, or intentionally set fires in northeast Santiago. Specifically, since 1920, there have been only 5,105 false or unjustified alarms (an average of just over 66 per year), and only 4,410 intentionally set fires (an average of just over 57 per year). Only a minuscule portion of intentionally set fires have been arson, according to fire department officials (10).

Since regulation began in 1929, false alarms have increased 7.2 times, unjustified alarms 158.5 times, and intentional fires have increased 163.0 times. However, it is most impressive to consider the remarkable rise in these problems starting in 1960, and especially after 1980. Perhaps this rise can be explained by the expanded use of fire prevention technology, which creates more opportunities for equipment malfunction. Or it could be due to the political unrest in Chile during the period, or even its economic expansion.

Nevertheless, it seems clear that there must have been growing cultural and technological inefficiencies which increased the costs of fire fighting. At present, 21.2 percent of all alarms are the result of false alarms, unjustified alarms, and intentionally set fires. This figure was only 4.4 percent when fire regulation began in 1929, and only 6.4 percent in 1969, just before President Allende took power. Another thing that is clear is that arson is neither a major cause of building fires nor need it be an important concern of building and fire safety regulation in Santiago, in spite of the fact that there is much popular concern about arson in the United States.

Structural fires

The CBS data set may be used to evaluate how effective fire safety regulation has been at reducing fires during the regulated period. In 1929, the Chilean government identified a problem and sought to improve safety conditions for the public benefit by legislation. It is not entirely clear how a simple decree 3/4 putting words on paper 3/4 could improve fire safety conditions. However, it is not the task of public policy research to scrutinize the means or practice of regulating but rather to examine the causes and consequences of it.

As with Baltimore (until perhaps recently), there is little evidence to support the notion that regulation in northeastern Santiago has been effective in reducing the number of fires. Figure 11 shows the number of structural fires per capita in northeastern Santiago for the period 1898-1997. Note that bar data represent real numbers or best counts. In years where the exact number of structural fires was not known, there is a line which gives the second method estimate, while the bar shows the first method estimate (best count). For most of the uncertain years, there is little difference between the two estimates. The data from 1914-1919 are simple linear trend estimates. The 1898-1913 data are for total fires, all of which are assumed to be structural (and thus probably overstate the number of structural fires in those years).

Figure 11: Structural fires in northeastern Santiago.
Figure 11: Structural fires in northeastern Santiago.

Per capita structural fires have increased 45.4 times since 1898 in northeastern Santiago, 9.1 times since 1920, and 8.9 times since 1929. By way of comparison, per capita structural fires in Baltimore for these periods (up to 1994) rose 6.6, 4.7, and 2.6 times. Northeastern Santiago's nearly nine-fold increase may have been less than the over fifteen-fold increase in Baltimore during the regulated period through 1994, but the difference between them might be explained by a variety of things including: cultural and community differences, length of the time series (135 years in Baltimore but only 68 years in northeastern Santiago), or institutional differences, viz., that CBS is a private enterprise run in part by private resources whereas the fire protective services in Baltimore are funded entirely by government.

The increase in structural fires in northeastern Santiago has clearly been more dramatic than in Baltimore. Corresponding with this increase, the building code has grown from 2 pages to 616 pages -- without taking into consideration the additional pages of codes faced by builders from each comuna's local regulations. For example, there are over 68 pages in the comuna of Santiago, 50 in Las Condes, and 102 throughout Providencia (11). There are presently hundreds of bureaucrats involved directly or indirectly in building regulation in northeastern Santiago. Hence, there has been a lot more regulation but also many more fires.

Of course, it is impossible to say what would have happened without the fire safety regulation. Perhaps there would have been even more fires. We also must take into consideration the increase in density, although it seems unlikely that doubling density would alone cause the number of fires per capita to increase nearly nine times. Yet, it might partly explain the increase in the number of building fires per capita.

Southwick and Butler reported that "higher density cities tend to have greater fire losses" but that "larger cities, ceteris paribus, have somewhat lower fire loss rates" than smaller cities (Southwick and Butler 1985, pp. 1061-1062). However, the fact that fire losses are positively correlated with population density does necessarily mean that the number of structural fires with population density. Intuitively, it makes sense that loses from any single fire would be greater when density is higher.

Moreover, given the technological and transportation improvements since 1929 which have improved fire protection services, it seems unlikely that the density change would be the major determinant of the very large increases in the number of structural fires per capita. Nevertheless, we can say this much: there is no evidence that fire safety has been improved by building safety regulation in northeastern Santiago.

page 18


USC Seal

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

ISSN 1548-6036

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