“The amount of activity the Portsmouth Police Department (PPD) handles is very low, and does not justify either the current levels of staffing or overtime.”
Mayor Steve Marchand – 2005
TO: CITY COUNCILORS
FR: STEVE MARCHAND
DT: MAY 15, 2005
RE: POLICE OVERTIME & FY06 BUDGET
During the Police Department work session of Thursday, May 12, the subject of overtime was central to much of the Q&A. The format of the session was not conducive to discussing data much beyond the “how do you use it” level of analysis, but it seems fair to say that one’s position on what the Police Department’s FY06 budget should look like depends largely on one’s analysis of the department’s staffing, workload, and overtime levels. It is my guess that there will be relatively quick and harmonious decisions made on the Fire, School, and Municipal budgets, and that the fundamental question on the Council will concern the Police budget.
Although I have spent far more time this year analyzing the school budget, I’ve also spent a fair amount of time on the police budget, and I’ve come to a very strongly-held conclusion. There is no department in the City with more potential for delivering more value per dollar – and freeing up more dollars to either deliver more services or restrain tax increases – than the Police Department. This is not to imply that other departments don’t have real opportunities to work smarter, or to better match spending with citizens’ needs and priorities. It’s just that the Police Department has real opportunities to save substantial amounts of money immediately – simply by analyzing data, and without cutting personnel or sacrificing service levels.
There are at least two reasons overtime can be cut significantly in FY06.
1. The amount of activity the Portsmouth Police Department (PPD) handles is very low, and does not justify either the current levels of staffing or overtime.
The basis for all law enforcement staffing decisions is Calls for Service (CFS). CFS are those police activities that are the direct result of calls made to dispatch, whether the call involves a barking dog or an assault and battery. CFS do not include such activities as speeding violations or neighborhood policing, which are classified as Self-Initiated Activity (SIA). The time spent on CFS – that is, the time between when the call is received by dispatch until the time the officer(s) leave the scene – is considered “committed” time.
The established benchmark for the percentage of all patrolmen’s time that is spent “committed” – that is, answering CFS – is 30%. If a shift’s committed time is significantly exceeding 30%, it means the shift probably doesn’t have enough manpower. If it is markedly below 30%, the opposite applies. (As to what occurs during the other 70% of the shift’s time, this includes neighborhood policing, non-directed patrol, other programming (child car seat programs), and the paperwork associated with CFS).
All police departments have the ability to keep track of CFS, if for no other reason than dispatchers log all calls coming into the department. What separates departments, however, is how they use CFS, or for that matter, if they use CFS to determine how to deploy resources.
To determine the PPD’s level of committed time, I took a representative sample of all CFS between April 2003 and January 2004 in Portsmouth. Specifically, I recorded all CFS during the first seven days of every third month (April, July, and October 2003, and January 2004), and recorded during which day of the week, and what hour of the day, each CFS occurred. This is done to ensure a sample that reflects all four seasons, all days of the week, all times of day, and all types of CFS. Then, the numbers are annualized, and you can tell how many CFS the PPD can expect during a typical hour/shift/day.
The results of that data collection are on the next page. It basically answers the question, “How many CFS, on average, are occurring per hour, on each day of the week?” For example, according to the data, there are an average of 1.25 CFS on Sundays, between midnight and 1 AM, and 2.75 CFS on Sundays, between 1 AM and 2 AM.
There are a number of observations that can be made from this analysis, including:
- With the exception of Sundays, which have the lowest number of CFS of all days, and Fridays, which have the highest number of CFS, the other five days appear to have very level numbers of CFS each day. This is not surprising, when you consider that the two biggest drivers of CFS are automobile activity and alcohol-related activity. Sunday, the slowest day for CFS, is the only day of the week with neither a typical rush hour nor nightlife. Friday, the busiest day for CFS, is the only day with both a typical rush hour and nightlife. Monday through Thursday have rush hours, but little nightlife, and Saturday has nightlife, but no rush hour. Again, this is a typical CFS pattern, and not surprising.
- One of the striking results of this analysis is that, regardless of the day of the week, or time of day, there are simply not a lot of CFS in the City of Portsmouth. In this sample, there are no one-hour time blocks where the CFS per hour reach 4.0. In other words, on average, there is no hour of the day, in a given week, where the PPD receives 4 CFS. While there are peaks of activity in a relative sense (some times are busier than others), it is hard to say that there are times of day with high CFS workloads.
- One time of day that appears to have higher general levels of CFS is in the 11 AM to 3 PM block. It is probable that this is related to lunch hour traffic, but an analysis of where these CFS are occurring might provide insight as to some of these CFS could be prevented. For example, if it is found that many of these calls occur in the downtown business district, directed police presence at key intersections where accidents are occurring could cut down on accidents and other CFS. Not only would this lower the chance of property damage or personal injury, it also would free up police resources (namely, patrol personnel) to conduct other activities. Also, because the end of this block of time also represents the end of the daytime patrol shift, CFS which occur in the last hour or two of the shift tend to require patrol overtime, as they must complete administrative work prior to leaving for the day. Every CFS avoided by targeting presence to likely areas of activity between 1 and 3 PM likely cuts down some of the costs associated with overtime.
More directly related to the issue of reducing overtime costs is using these data to compare CFS activity by shift. What follows is the number of CFS per shift, for each day of the week. This chart uses the same data as the overall CFS chart.
Call for Service, Portsmouth Police Department, by Patrol Shift and Day of Week
Based on sample of CFS in April, July, October 2003, and January 2004
|11 PM-7 AM||9.25||3.75||5.5||3.75||7.25||8||9.75|
|7 AM-3 PM||11.75||17.25||18.75||13||17||15||17.75|
|3 PM-11 PM||9.5||14||12.5||16.75||12.5||16.5||9|
In other words, the 7 AM-3 PM shift receives, on average, 11.75 CFS, for the entire shift. In the case of the overnight shifts, the average CFS represents the last hour of the previous night, followed by the 12 AM-7 AM portion of the day in question (for example, the 11 PM-7 AM shift under “Sunday” is Saturday night, from 11 PM to midnight, as well as the midnight to 7 AM on Sunday). The following observations can be made from these data:
- A well-run department should have about 30% of its man-hours dedicated to CFS, or “committed”. According to the Police Department’s comments during the recent work session, the overnight shift typically has six patrol officers, for a total of 48 “man hours”, while the daytime shift has 10 (for a total of 80 man hours), and the late day has eight (for a total of 64 man hours).
- While the national average for committed time per CFS is between 20 and 30 minutes, the Department said at the same work session that its average CFS committed time is only 15 minutes – an unusually low number. The average CFS time is just that – the average of the calls that take five minutes (barking dog, somebody blowing through a stop sign in another part of town, etc.) and the calls that take considerable time (a significant car accident, assault and battery, etc.). If the average time in Portsmouth is actually 15 minutes, this indicates a simple lack of time-consuming calls. For the purposes of analysis below, the average time per CFS has been increased from 15 minutes to 30 minutes, to provide benefit of the doubt to the department.
- If there are 9.25 CFS on a Sunday overnight shift, at 30 minutes per CFS, this would mean the committed time for the shift is 277.5 minutes, on average, or 4.63 hours. If the overnight shift has six patrol officers, this would mean there are 48 man hours available. 4.63 committed hours out of 48 man hours means the shift has a committed rate of 9.65%. Again, this is assuming each CFS takes an average of 30 minutes – double what the department themselves estimates. This committed rate is less than one-third the desired figure, and suggests the shift would provide the same level of service with a fraction of the manpower currently available.
- The example above actually provides one of the busier overnight shifts of the week. Here is the same chart as above, providing the average committed time rates for each shift, each day of the week:
Committed Time Rate, Portsmouth Police Department, by Patrol Shift and Day of Week
Based on sample of CFS in April, July, October 2003, and January 2004
|11 PM-7 AM||9.64%||3.91%||5.73%||3.91%||7.55%||8.33%||10.16%|
|7 AM-3 PM||9.18%||13.48%||14.65%||10.16%||13.28%||11.72%||13.87%|
|3 PM-11 PM||5.94%||8.75%||7.81%||10.47%||7.81%||10.31%||5.63%|
- The busiest shifts, in terms of CFS committed time, are the daytime shifts on Tuesday, Saturday, and Monday. That said, no shift in the entire week, on average, reaches even half of the 30% desired committed level – and this is assuming each CFS takes twice the time the department says it does. If the department’s figure of 15 minutes per CFS is correct, these numbers are cut in half, across the board.
- The slowest shifts, in terms of CFS committed time, are the Monday and Wednesday overnight shifts, followed by the Saturday and Sunday late day shifts, and the Tuesday overnight shift. This is a key concept, and one many miss. While many of the slowest shifts, in terms of committed percentage, are those shifts one would expect to have the fewest CFS (overnights during the week), this does not suggest that the slowest shifts should also have the lowest percentages. What it means is that the department should have a fraction of the patrol size overnight than it has during other shifts, because the committed percentage should be roughly uniform, across all shifts.
- To put these figures in another light, an average patrolman spending 30% of their shift on committed activities would be spending about 2.5 hours per shift on CFS. At 30 minutes per CFS, this would allow for about five CFS per shift. The busiest shift, in terms of CFS, in the sample was the Tuesday daytime shift, at 18.75 CFS per shift. This would represent fewer than four patrolmen performing at 30% committed time – at a time when there are eight regularly scheduled. This example represents the busiest shift in the department. Some of the slowest shifts (weeknight overnight shifts) would require fewer than one patrolman to keep staffing in line with CFS workloads. No overnight shift requires more than two patrolman, on average.
The bottom line in all of this analysis is that it is difficult to see more than a handful of occasions during the year when the PPD should need any carryover (that is, keeping a patrolman for multiple shifts) or callback (that is, calling somebody in from their home) overtime – even assuming the average CFS committed time is double what the department itself says. There is simply not the activity in the city to justify anything close to the level of staffing – much less the need to augment that staffing with overtime – currently being deployed.
2. The current deployment of detectives is extremely inefficient, with a staffing policy that is both unnecessarily expensive, and unresponsive to citizens needs.
Currently, there are 14 detectives on the PPD force. Nine of them are general detectives – that is, they investigate most of the crimes one might commonly link with detectives, such as vandalism, theft, and white-collar crime. There are also five Child and Family Service (CFS) detectives, who specialize in dealing with cases involving children, domestic disputes, etc. The School Resource Officers (SROs) are included in the CFS unit.
In my former work as a government auditing consultant for Maximus, I probably spent as much time interviewing detectives, and reviewing their case logs, as anything I did. I have not reviewed either the number or the nature of the cases being handled by the investigative unit, and that is not my primary purpose for analysis this year. The only comment I would make in this regard is that, based on the low levels of CFS in Portsmouth, it is quite possible that the workload indicators for the investigative unit would show a similar inefficiency.
That said, the immediate question deals with the deployment of the investigative unit. All 14 of the investigators are scheduled for the daytime (7 AM to 3 PM) shift – and any work done outside of those hours is considered overtime. While this is not unprecedented, it is also highly inefficient, in that many of those who would be interviewed are commonly unavailable during the day. It is also the case that certain tasks in investigations (calls to other time zones, online research, and simple review of physical documents) are most efficiently done at times when there is less activity inside the department.
By shifting at least one Child and Family Service detective to the late day shift, and shifting approximately three general detectives to the late day shift, at least three desirable outcomes would be achieved:
1. Detective-related overtime would decrease, as more of the types of activities that occur after 3 PM (interviews with adults with daytime hours, interviews with children after school, etc.) would be performed by detectives on their regular shift.
2. The Department would get more value per dollar, as the overtime currently being earned by detectives is at a rate of time-and-a-half. Considering the higher rate of pay for overtime, it is reasonable to say that wise deployment could reduce overtime enough to pay for the additional dispatcher they could probably use (I haven’t analyzed dispatch workload activity, but it appears possible that there is a need for an additional dispatcher).
3. The quality of service to the community would actually increase, despite spending fewer dollars on investigative services. The current deployment strategy encourages a pattern of investigators calling empty homes during the day, followed by adults coming home late in the day. The adult hears the message, calls back the investigator, and leaves a message of their own. Then, the investigator gets the message, calls the adult in the evening, and incurs overtime to have a conversation with the adult during the late afternoon or early evening. Besides the overtime costs, the bigger problem is the 24-hour delay in reaching the interviewee. By deploying personnel based on activity levels, we can improve customer service, and save overtime costs, to boot.