Congestion in Delivery of Emergency Services in Urban Areas: The Case of Police Response to Burglar Alarms

Erwin A. Blackstone, Simon Hakim, and Uriel Spiegel[1]

September 5, 2001

Congestion in Delivery of Emergency Services in Urban

Areas: The Case of Police Response to Burglar Alarms

ABSTRACT

Urban governments and in particular large cities face an increasing demand for emergency services. These services have both public and private attributes. If the call for service is real then public involvement but not necessarily provision is justified. However, in case of a false incident it is a private good, and the efficient solution involves market provision. Whether the service is indeed public is unknown until the service is actually delivered. This attribute of this newly defined public good suggests a public private partnership to deliver the good in a socially efficient manner. This solution differs from existing practices.

The emergency service analyzed in this paper is police response to burglar alarms. Such response comprises twenty percent of police efforts. Further, 94 to 99 percent of the burglar alarm calls for service are false. They are thus essentially a private service. This paper suggests an initial private response to an activation followed by police response if an actual burglary is evident. The private response portion in most areas can be provided competitively.

1. Introduction

Urban Governments throughout both North America and Western Europe, including in particular Great Britain, experience congestion on both their 911 systems and in police response to emergencies. Government provides most emergency services for equity and/or efficiency reasons. Society chooses for equity reasons to make rescue from life threatening situations generally available so that the poor cannot be denied assistance. Market prices would often be too high to be afforded by the poor. In the case of ambulance, hospital emergency, or Coast Guard services the threat is mainly to human life while in the response to burglar alarms it is invasion of property rights and possibly violent incidents.1 In the case of fire protection the public interest lies in preventing the spread of fire, a negative externality that threatens both lives and property. Government has usually subsidized these services so prices are below market levels in order to make the service affordable to all. As a consequence, individuals may abuse the system and claim they have an emergency in order to obtain the public service.

This paper discusses the problem of congestion in 911 emergency services that is prevalent in large metropolitan areas. Emergency services often have both private and public attributes. We introduce a new type of public good whose attributes dictate a mixed delivery by public private partnerships. We show that society’s welfare could be improved by shifting some production and financing responsibilities to the private sector where competition for delivery prevails; in some cases government can effectively compete with private providers.

The analysis is applied to police response to burglar alarms. Response to false alarms causes congestion on the 911 lines, reduces police response to real and serious events, imposes cost on police, and increases the threat of accidents from rapid police response. The false alarm problem is particularly acute in large cities where the police are already burdened with much crime and where the absolute number and growth of alarm ownership yields a large number of false activations. In Philadelphia, for example, in 2000 only 3,022 out of 140,604 alarm responses involved actual or attempted burglaries. For panic alarms, 430 responses out of 18,247 were bona fide robberies. On the average each Philadelphia system was activated twice compared to 1.3 for the U.S. as a whole, suggesting that the false alarm problem is most serious in large cities. Further, other crime tends to be more prevalent in large cities, implying that their real cost for alarm response is probably greater.

To determine an efficient provision of response to requests for emergency services, including alarms, requires understanding the nature of the service. Since government delivers the service, it is helpful to discuss the theoretical evolvement of public goods. Samuelson (1954) was first to identify a “pure” public good. Such a good may not be produced at all or in non-optimal amount under free market conditions. However, the good can provide net social benefits and only government which represents the public interest would supply the optimal quantity. It is very costly or even impossible to exclude any one from consuming a public good, and each and every person consumes the full amount of the output. Without government forcing all to share the cost, each person would have a strong motive in becoming a “free rider” or to pay less than the socially optimum amount. Thus, there is essentially no alternative but for government to take responsibility for the supply of public goods. Clearly, however, government does not need to produce the good and could let that be done under competitive market conditions.

The assumptions necessary to classify a good as a pure public good are seldom completely met in reality. For pure public goods the size of the interacting group is the entire society, and the entire supply is commonly consumed. If either or both the requirements of non-exclusion and non-rivalry fail to be met then it falls into the general category of an impure public good. These impure public goods incorporate the notion of congestion cost or excludable benefits. Unlike pure public goods, a larger number of consumers may cause congestion in the consumption of impure public goods. Examples include swimming pools, tennis clubs, golf courses, and highways. (Cornes, and Sandler, 1986: 4).

Buchanan (1965) and Olsen (1965) elaborated on Samuelson’s work by defining a more narrow type of impure public good that is frequently found in local communities. In the case of Buchanan’s Club good, the entire supply is commonly consumed; however, the size of the interacting group is small. The level of individual benefits decline with the size of the population and the distance from the source of supply.

Another “variant” public good that is similar to a Club good is the local public good. It assumes congestion but does not incorporate excludability for the population within the jurisdictional level. Examples include local schools, public libraries, or public parks. Police patrol that provides an “umbrella” of security to the community can also be classified as a local pure public good.

Buchanan’s and Olsen’s definition of Club good assumes that consumers travel to a defined geographical area where the service is provided. However, a “variant” local public good can also be provided at the consumers’ own location as long as the two assumptions of excludability and congestion are satisfied. For example, patrol officers educating children at their schools about the dangers of illegal drugs are still providing a local public good. This is a public responsibility since significant externalities exist.

Local governments often provide emergency services because of life threatening conditions and/or a significant level of externalities. In case of a major disaster, like an earthquake, emergency services need to be in place in order to serve the general population. Emergency services include fire protection, police response to alarms, stray animals, gas odor, and ambulance services. The Coast Guard, a federal agency, provides quasi-local services to stranded boaters. These are all local rather than pure public goods since congestion occurs, and residents in other adjacent localities are excluded from enjoying the services. Government provides these services because of their significant externalities and potential life threatening conditions.

Emergency services are distinctively different in one important aspect from local and “variant” public goods. The output of both local and the “variant” local public goods is a priori certain at a high probability level. Emergency services have a common attribute; it is unclear whether a real emergency exists at the time service is requested and the emergency crew is sent out. Only when the service is actually rendered does its “emergency” nature become known. For example, poor people without medical insurance often arrive for emergency treatment at expensive hospital emergency rooms where service can normally not be denied even when it turns out to be a non-emergency situation. Indeed the service is often delivered before its emergency status is known. Public ambulances are also often dispatched for what turns out to be non-emergency events. The case of police response to burglar alarms is another example. When police are dispatched it is a priori unclear whether a real break-in has occurred. Only after the officers actually provide the service is it known whether an actual break-in is in progress or has occurred. Hence, in all these cases of emergency services, the probability of a real event is less than one. For example, in the case of police, ex ante response to false alarms will be shown to be a public good at a probability level of at most 6 percent.

In case of a real event, public intervention can be justified. However, in case of non-real event, public financing or intervention is unwarranted. When a real break-in occurs, the public interest requires that police attempt to catch the burglar. Apprehending burglars diminishes the pool of burglars and reduces the probability that others will become victims of burglary. Apprehension also has a deterring effect by raising the cost of criminal activity, and thereby may even reduce the future supply of burglars.

When a false alarm occurs it is a private good and government intervention is unjustified. If a water pipe in the home breaks, it is obviously an emergency; however public intervention is not warranted. Similarly, when an alarm is falsely activated, no one else in the community derives any benefit from the response to the false alarm.

2. The false alarm problem

False alarms are a significant problem in all emergency services. For example, 53 percent of 911 calls in Atlanta during 1997 were of a non-emergency nature and added to the congestion on the system (FCC Docket on Assignment of 211: 7). During the year 2000, Philadelphia police reported that 96 percent of their responses to fire alarms, 97 percent for robbery alarms, and 75 percent of medical alarms were of a non-event nature.2

We chose police response to alarms as a case study for all emergency services. Alarm systems are purchased from alarm dealers. Dealers sign up the alarm owner for monthly monitoring by a central station. Large dealers usually have their own central station while smaller dealers contract out monitoring to central stations that serve customers of many dealers and enjoy economies of scale.

When an activation occurs, a signal is sent to a central or monitoring station. The central station is supposed to verify whether an intrusion or attempted intrusion occurred. In case of a suspected real event, the protocol is for the central station to request police response. In most situations two officers respond in one or two vehicles. More than one officer is required because of the possibility of a violent confrontation if the burglar is on site. Further, most police departments will not accept a central station’s cancellation of the request for police response once the police have been dispatched.

The problem is that 94 to 99 percent of all activations are false. For example, DeKalb Georgia police in 2000 found that only 39 out of over 144,000 alarm calls were actual or attempted burglaries (SDM, January 2001:51). In Seattle Washington 97.5 percent of 30,000 police responses to alarm were false and only 40 burglars were actually apprehended. Chicago Police annually respond to over 300,000 activations, 98 percent of which are false. The corresponding figures for West Palm Beach, Florida are 66,000 and 99 percent (Kern, November 2000: 43).

Alarm ownership increases annually by eight to ten percent, yielding an almost identical rise in false activations. Seventy-six percent of false alarms are caused by subscriber error, ten percent by equipment malfunction, and the remaining fourteen percent by weather or telephone problems. On the average, each system activates falsely 1.3 times a year. The commercial rates are three times the residential rate. Banks, schools, and municipal facilities falsely activate seven to ten times the residential rate (AIREF, 1999: 10).

The cost of responding to false activations is substantial. The cost includes the value of the time spent by one or two officers in one or two vehicles. Additional cost items include the time spent by the dispatchers, vehicle costs, and the expected cost of accidents attributed to the response required under mainly emergency situations. Indeed, police officers in Topeka, Kansas; King County, Washington; and Calgary, Canada died in the late 1990s in highway accidents while responding to false alarms. Administrative costs include supervision and data management by police and possible court costs. For example, Seattle Police Department’s calculation adjusted for overhead cost yields a $52 average cost per response in year 2000.3 Further, communications costs of false alarms for the 911 systems are substantial. For example, Seattle, Washington in 2000 incurred $303,237 just to process 911 alarm activations calls.4 Table 1 illustrates the cost of false alarms for a sample of communities.

The cost of patrol officers is a partial indication of the significance of the false alarm problem. Portland, Oregon estimated in 1998 that the equivalent of 45 full time officers responded to false activations. Winston-Salem, North Carolina utilizes the equivalent of 7 officers, Philadelphia 52, and Los Angeles 100. Nationwide, most communities claim that ten to twenty percent of police resources are devoted to false alarm activations. In fact, false alarm calls for police response comprise ten to twenty percent of all calls; for example, in 1998 in Ft. Lauderdale the percentage was 14.7; in

Redmond, Washington 12.4; and Coral Gables, Florida 10.3, and in Palm Beach County, Florida the percentage was especially high, 23.6. In West Palm Beach, Florida, alarm calls comprised as much as 25 percent of patrol duties.

Not surprisingly, police responding to mainly false burglar alarms apprehend relatively few burglars. Consequently the cost per arrested burglar is high. In Seattle as a result of alarm activations in year 2000 the cost was $38,500 for each of the 40 apprehended burglars. Such a high per burglar cost raises the issue that the social benefits of alarm response may well be less than the costs. However, some police departments set a low priority for patrol response to alarms. This policy reduces the chance of apprehending burglars.

The cost of false alarms rests on police departments and over time the problem is becoming more acute since police budgets rise at less than three percent annually while false alarm response rises by almost ten percent. Hence, police need to divert patrol officers from security activities that benefit the community, a public good, to serve alarm owners who falsely activate their systems, which we argue, is a private service. Furthermore, officers who are trained to handle crisis situations are expecting to confront a burglar and instead deal with a non- event.

Table 1

Calculated Cost of False Alarms

CITY	YEAR	NUMBER OF FALSE ACTIVATIONS	CALCULATED COST	TOTAL COST	REFERENCE
(1)	(2)	(3)	(4)	(5)	(6)
Nationwide	2000	36 m	$50	$1.86+	Security Sales, 2001, Vol. 22 No. 12 P.8.
Philadelphia, PA	1995	147, 074	28	4.3 m	Hakim and Blackstone, 1996 Percent false 97.4%.
Dade County, Fl.	1995	121, 717	24	2.9 m	Hakim and Blackstone, 1996.
Reno, Nv	1995	11,185	72	0.8 m	Hakim and Blackstone, 1996
Phoenix, Az	1995	47,075	73	3.4 m	Hakim and Blackstone, 1996.
Salt Lake City,	1999	8,213	60	0.5 m	Material supplied by Salt Lake City.

8a

3. Practices to solve the problem

Police departments across North America, Great Britain, and elsewhere struggle with the false alarm problem. Municipal governments have experimented with their ordinances to try to reduce the problem. Ordinances have been adopted that increase and escalate fines for repeat activators, cease response after a certain number of false activations in a calendar year, require education of chronic false activators, mandate registration of alarm systems, and require dispatching calls to use 900 lines.

Some ordinances have even deceased fines. No study has been conducted to evaluate the relative effectiveness of the numerous policies. Such research would lead to a social cost-benefit calculation for each policy/practice to determine the most appropriate alarm ordinance.

A common practice is to increase fines and also to introduce an escalating fine structure for multiple false alarms during a given time period, typically one year. Beset with a large number of false alarms, Clearwater, Florida in 2000 increased its false alarm fine from $30 to $50. New Orleans assesses a fine of $25 for the sixth false alarm in a year with increases to $75 for the fifteenth. Itasca, Illinois charges $75 for the fourth and escalates to $750 for the tenth. Montgomery County, Maryland assesses as high a fine as $4000.

Communities normally allow some free false activations before fines are assessed. The rationale is that alarm owners ought not to be penalized for equipment malfunctions or user error unless the problem occurs often. Montgomery County, Maryland is fairly typical in its allowing three free false alarms in a calendar year. New Orleans allows five such activations. However, Seattle, Washington charges $50 from the first false alarm.

Some experimentation with ordinances to reduce the false alarm problem has involved lowering false alarm fines. Beaumont, Texas stopped charging for false alarm response in December 1977 and instead substituted a $50 charge for reinstatement, which meant that police officers would resume response. Under the new ordinance, officers would respond to eight alarm activations in a year at no charge but then would cease response until a $50 reinstatement fee was paid. An additional false alarm would require a $100 reinstatement fee to once again obtain police response. Under the former ordinance, a fine of $50 was assessed for both the seventh and eighth false activations. The new ordinance has led to an increase in the number of false alarms, a reduction in the revenue obtained from alarm response, and an increase in the number of locations on no- response status from 19 in 1998 to 76 in 1999 (Zarazua, 2000). Similarly, Bartow, Florida in November 2000 was considering increasing the number of free alarms from two to three and reducing the maximum fine from $500 to $250. It was also considering the elimination of fines for those who repaired their malfunctioning alarm systems (Kern, 2000: 43).

Municipalities sometimes assess different false alarm fines for residential and commercial establishments. Commercial establishments have a higher propensity for false alarms, have usually greater ability to pay, and are often required by insurance companies to purchase and operate an alarm system. Commercial establishments are thus more likely than households to pay fines, even if they are high, to maintain continued police response. Irvine, California, for example, fines commercial establishments twice as much as residences. The maximum fines in Montgomery County, Maryland for residences and commercial establishments are, respectively, $1000 and $4000.

Some ordinances are confusing and complicated and make compliance difficult. Confusing ordinances diminish the deterrent effect for alarm owners and confuse dealers who typically operate in many localities. For example, in Vancouver, British Columbia charges for false alarms begin with the fourth, which is $75 for residential, $125 for commercial, and $250 for large commercial users. The fifth through the seventh in a calendar year are then free, but the eighth false alarm is twice the fourth or $250 for commercial alarm owners. Obviously, such a cumbersome structure has limited deterrence effect and serves to confuse all involved. Palm Beach County, Florida charges different registration fees for alarm owners depending upon the total false alarm experience of the dealer from whom the alarm system was obtained. Alarm consumers would have to base their purchase decision on the seemingly unknowable future false alarm experiences of the dealer’s other customers. The intent obviously is to provide the dealer with a strong incentive to reduce false alarms. However, such a practice is unfair to careful alarm owners who happened to contract with a dealer that serves other repeat activators.

Registration of alarm systems is also used by municipalities to reduce the number of false alarms and to pay the fixed cost of maintaining the alarm unit, usually within the police department. The normal practice is for the municipality to charge an annual registration/permit fee of between $10 and $50 for the privilege of having an alarm system. Southlake, Texas, for example, has an ordinance that requires annual permits of $10 for residences and $20 for commercial establishments but permits are not required for residents at least 65 years old (Kern, 2000: 43). Permits can be revoked or a policy of no response adopted after five false alarms in a calendar year. Nassau County New York requires a permit in order to have police response to an alarm (SDM, 2000: 28). Port Hueneme, California adopted in 2000 a differential fee structure for permits for residential and commercial alarm systems. The city expects to collect sufficient revenue to increase the number of patrol officers and to reduce the number of false alarms (SDM 2000: 28). Similar to the fine situation some cities like Chicago require only businesses to obtain alarm permits.

Registration of alarms involves substantial resources and is often difficult to achieve. In Redmond, Washington in 2000 only 1200 alarm systems were registered out of an estimated total of 12,000 (SDM, 2000: 28). Registration in Redmond has been unsuccessful even though the police will not respond to activations at unregistered locations.

Non-response to repeat alarm activations is also used to reduce their number. New Orleans police will not respond to burglar alarms for repeat violators, defined as more than 20 in a year. In 2000, 280 locations with alarms were on the Department’s list for no response (Philbin, 2000: Metro, p.2). Seattle, Washington, which suspends response after six in a twelve-month period, typically has between 8 and 50 locations on a no response policy (AIREF 1999: 88). Irvine, California suspends response after the seventh false activation in a year. Kenner, Louisiana police do not respond to business alarms during daily peak false alarm periods. They also do not respond to business activations during very intense thunderstorms, which typically cause sensitive systems to activate falsely (Hyman, 1999: Metro, p. B1).

Non-response is a serious problem for the industry, since alarm owners will either stop using their systems and consequently stop paying for monitoring of their systems and/or non owners will be reluctant to purchase a system. High fines could lead to the same result. Toronto, Ontario in 1996 started assessing high ($77) fines from the first false activation and ceased response after the fourth. Private response quickly developed to cover over 90 percent of the systems in Toronto.

Education of consumers and sometimes dealers, is also used to reduce the number of false alarms. The rationale is that user error accounts for 76 percent of false alarms and also that about 20 percent of users account for 80 percent of false alarms. West Palm Beach, Florida residents on their third false alarm can pay a $250 fine or attend alarm school. The one-hour class is designed to show the attendees the magnitude of the false alarm problem and to provide repeat activators with suggestions on how to reduce such activations (Kern, 2000: 43). New Orleans police have had a seven-hour class for alarm companies deigned to reduce their false alarms. The Department also loans videos to alarm system owners on how to reduce false alarms.

The Fort Lauderdale, Florida police have been especially active in educating consumers and dealers. The police department’s alarm coordinator had a meeting to which 250 alarm companies were invited while only 23 sent representatives. The purpose of the meeting was to focus effort on those alarm owners who had the largest number of false alarms in 1998. The attendance of less than ten percent of the alarm companies indicates lack of interest for the program. However, one alarm company with a large percentage of the false alarms in 1998 did attend and even assigned a technician to work with the police department’s alarm coordinator. The coordinator, along with an alarm industry representative, visited alarm owners who had the greatest number of false alarms. One automobile dealership reduced its false alarms from 100 per year to none. Overall the education efforts have reportedly reduced false dispatches by 8 percent (AIREF, 1999 48).

The alarm industry established two programs to work with police departments to reduce false alarms. The above Fort Lauderdale effort was part of the second program, the Model States Program, towards which the industry contributed over $900,000. The alarm industry also developed and provided municipal officials at no charge software using dealers’ databases with their current information on alarm owners and their addresses. This information is thought to be helpful to the police in updating their database to better conduct their enforcement activities.

An effort directed at dealers occurs in Toronto Ontario where police impose the false alarm fines on the alarm dealers who then have to charge their customers or absorb the fines themselves. This reduces the transactions costs for the police who are now involved with only 70 companies instead of thousands of alarm owners. The dealers are challenging the police department practice in court. Making dealers responsible for collecting fines and then suspending police response to all the customers of the non-compliant dealer is strongly opposed by the industry.

Another practice that has also been strongly opposed by the industry is requiring calls for police response to alarm activations to be made on 900 lines. Riverside County, California enacted ordinance number 753 which mandated such a policy. The ordinance was to become effective on August 5, 1996. All calls had to be made on the special 900 lines with an automatic $5 charge to the central station. The alarm industry secured an injunction preventing enforcement of the ordinance on the ground that requiring crime victims to pay to report a crime and request assistance violated the due process and equal protection provisions of the Constitution (California Alarm Association, 1996).

Finally, the alarm industry as represented in its Model States Report supports allowing physical response by police to be cancelled if the central station learns after dispatching police but before police arrive at the location that the alarm is indeed false. It further supports strong enforcement of the alarm ordinances, registration of alarms, and reasonable and escalating fines. It encourages suspension of response for non-payment of fines or for exceeding the maximum number of allowed false alarms, verification of all alarms by central stations, and education of alarm system owners.

4. Evaluation of Current practices

The above practices may produce a reduction in false alarms. For example, not responding to alarms for repeat activities obviously reduces false alarms. Targeting repeat activators for intensive education may work short term as well. However, such programs and policies may reduce the perceived and actual level of security, since owners may not arm their systems. Such punitive policies may not be socially efficient; the social cost of reduced security may be greater than the social benefit of reduced false alarms and the resulting cost of police response. These policies also may not account for the public good portion of alarm response, namely, the benefits of apprehending burglars.

The alarm industry claims that response to false activations is a normal police function, a public good. In essence, the industry disputes both the private nature of response to false activations and the related argument that such public police response is socially inefficient.

The basic problem is that no one except the police has an inherent interest in solving the problem. The alarm owner on the average falsely activates her system 1.3 times a year. Since most communities allow three false activations a year, alarm owners usually incur no charges.

Alarm dealers view police response as a gift to their business.

They sell a system, charge monthly fees for monitoring the alarm system, and when necessary, obtain police response that is essentially paid for by the general taxpayers. Dealers consider false activations to be an issue merely between the police and the customers. There is also little apparent interest by individual dealers to spend resources to solve their own customers’ and the communal problem. The latter is the regular “free rider” problem. Individual dealers view their private cost of reducing their customers’ false activations to be greater than their private benefit. Thus, dealers and the alarm associations contest police efforts to cease response, except for non-payment or for repeat activators. They also object to shifting responsibility for their customers’ fines to them or to the central stations. The alarm industry prefers that police rather than private companies respond because police response usually makes their product less expensive to customers and thus more appealing. Increased fines for more alarm system

owners will also reduce the purchase of new systems. It is quite ironic that private industry prefers government intervention in the marketplace rather than allow new forms of business, companies offering alarm response, to develop.

Central stations are supposed to verify that an actual intrusion took place and only then to dispatch to the police. However, suppose that a call is verified as false, but a real break-in did indeed occur. In such a case, the central station is liable for breaching its obligation and is likely to suffer negative publicity with perhaps sizable liability suits. Indeed, the central station bears no cost for dispatching police to a false activation while the cost could be high for not dispatching to a real event. Hence, the central station dispatches the police even when the probability of a break-in is extremely small. Sometimes the central station determines that the call was false after the signal was dispatched for police response. Central stations then attempt to cancel police response. However, police most often do not accept cancellation after they have been dispatched. Cancellation with its added calls itself adds substantial cost to the 911 system. For example, Seattle Washington incurred $97,000 in 2000 for processing cancelled alarm calls.

The alarm associations with a broader view of the just industry realize that the false activations will eventually mean deteriorating business for dealers and central stations. Indeed, the industry has tried to solve the problem. However, the associations as part of the industry would like the police to maintain their free service that benefits their constituents, and in any case the associations have limited influence on their members to curb false activations.

The intuitive practice of ceasing response is economically inappropriate. Suppose that a person has a habit of slamming the garage door and breaking it. As long as the person pays the bill, the repair company will send technicians to repair the damage. No company will deny response after a certain number of calls. The more calls, the better it is for the garage door repair company. In the same vein, there is no reason to deny response to false activations as long as the violator pays the full cost, where no one else is made worse off. The socially efficient solution requires that price reflect the social cost of response.

There is also no reason to cease response during peak alarm periods. The charge should simply be higher, reflecting the straining at capacity. After all, garage door companies still respond when business is booming as long as the customer pays the required charge. Similarly, the police should not decide arbitrarily when to provide services. The socially efficient solution requires response at all times as long as social costs are recovered.

Police often charge higher fees for commercial establishments than for residences, and they normally exempt municipal facilities from charges. Businesses pay higher fees because their demand for response to alarms is more inelastic than is the demand by households. In addition, business owners are often non-residents so that local politicians view them as a desirable source of revenue. Since prices are not cost based, price discrimination is occurring, a practice common under monopoly conditions and often not socially efficient.

Sometimes the public good aspect of alarm response is ignored. Toronto, Ontario police charge for all responses to burglar alarms including actual events (Hyman, 1999: Metro, p. 81). The probability of catching a burglar and benefiting the community means that the price in the event of an actual or attempted burglary in Toronto is in excess of the correct price. When an activation involves a real or attempted burglary, social efficiency and equity principles suggest that no charge should be imposed.