Data-driven government in Baltimore

The initiative

After hiring a small staff and housing them in City Hall, CitiStat was launched. CitiStat grew from covering one department when it launched in June 2000 to 16 departments in 2002. “O'Malley's new approach rested on the idea that real-time information could, and should, be used to underpin decision-making, prioritise scarce resources and accurately track the impact of policies and programmes. Called CitiStat, it began with crime reduction but was soon extended to other areas such as healthcare and education. Under the programmes, a 311 phone number was introduced for all city services, and then this was underpinned by performance measurement and a collaborative problem-solving approach.” [3]

CitiStat gathers data from around the city, which is then stored and analysed. “Information is gathered on an array of performance indicators, including response times for things like pothole abatement, trash collection, and snow removal, as well as the prevalence of problems such as illegal dumping, vacant buildings, and sewage overflows. This information is analysed with the assistance of computerised databases and geographic mapping to zero in on areas of underperformance. Managers from each city department then meet with the mayor's office every two weeks to answer questions about their results.” [5]

The challenge

When Martin O'Malley took over as city mayor in December 1999, “Baltimore had allowed herself to become the most addicted, the most violent, and therefore the most ‘abandoned' city in America." [1] In fact, Baltimore had the second-highest violent crime rate among the nation's 30 largest cities. “Baltimore's high crime levels were the catalyst for his ideas, as was the success of New York's police department in addressing systemic and long-running law and order issues. The so-called ‘broken windows' theory held that lax enforcement of laws against petty crime creates an atmosphere in which serious crime is more likely to take place.” [2] New York's success led to a revolution in performance-measured policing in cities and towns all across the United States - and one of those was Baltimore."

In addition, the city government suffered from widespread absenteeism in its various departments. For example, in the Department of Public Works one in seven employees failed to report to work every day on average. This absenteeism required other employees to pick up the slack, which produced high overtime costs and a huge burden on the city's finances. It also became clear that accurate and up-to-date data was lacking.

The public impact

CitiStat made a very positive impact on crime and other areas. “Over the course of the next 10 years, Baltimore achieved a bigger reduction in 'Part 1' crimes, such as homicide, rape, robbery and arson, than any other city in America - according to FBI data. Although the city's drop in crime rates mirrored a national trend, as other major cities also saw large drops, it is clear that substantial progress was made.” [6] The city lowered its crime rate by 14 percent during CitiStat's first year, and from 1999 to 2003, violent crime fell nearly 40 percent.

It also helped reduce problem of absenteeism, helping the city save around USD6 million in the first year after the launch.

Stakeholder engagement

CitiStat was the brainchild of Mayor O'Malley. The targets to be set for different city departments (one example of a target was a 48-hours pothole repair guarantee) were decided in collaboration with internal stakeholders, i.e., the different city departments. The other main internal stakeholder was the CitiStat office itself. However, some unions and frontline employees opposed the project, as the reduction in absenteeism would ultimately reduce the overtime pay of those who were present.

The main external stakeholders were the citizens of Baltimore, who were able to communicate their complaints, ideas and opinions via the 311 number and other means. There was also significant interest at a higher political level: “CitiStat also serves as a mechanism to monitor the implementation of the administration's most important policy and administrative initiatives and to ensure the coordination of responsive government for all stakeholders, including the federal and state government”. [7]

Political commitment

For Mayor O'Malley, the CitiStat programme was a city government priority, and the need for data-driven government was based on his experiences serving on the council when “‘nothing about it was really timely, accurate or transparent.' He set about making some changes.” [8]

He and his colleagues also wanted to apply other city's successes to Baltimore. "'We were watching NYPD Police Department dramatically improve its performance using CompStat. We were seeing the results and the same, we wanted for our city.'"

O'Malley attended a number of CitiStat meetings, demonstrating that the programme was of great importance to him. The data gathering “never would have happened, however, without commitment at the top … O'Malley placed top deputies in charge of presiding over review sessions, while sometimes attending sessions [himself]." [9]

Public confidence

Clarity of objectives

The objectives stated at the outset were clear and maintained throughout under the authority of the mayor:

• The CitiStat strategy remained effective, as everyone knew that the staff of CitiStat were always operating with the mayor's explicit personal authority and approval.
• The objectives were clearly measurable, as data was at the heart of the mayoral initiative. For example, the reduction of the city's rates of crime - 14 percent during CitiStat's first year - and within those of violent crime - 40 percent from 1999 to 2003.

Strength of evidence

The idea for the project was inspired by a similar existing data governance system, primarily used for fighting crime, New York City's CompStat. Much of the CompStat system was replicated in CitiStat, although its role was extended to other city departments in addition to the police department. Soon after his election, O'Malley worked with Jack Maple, the former NYPD deputy police commissioner for crime control strategies, to adapt CompStat to the Baltimore police context. O'Malley quickly decided the CompStat approach would be useful in managing other city departments.

Mayor O'Malley also based his focus on a data-driven approach on his knowledge of the inadequate levels of knowledge of the previous administration. “‘I had served on the council and, knowing the huge cost of our fleet of vehicles, I asked the outgoing director of public works how many vehicles the city owned and he said it was “one of those things we never really got on top of”. [10] It was something like 6,000 to 6,500 but that was city government as city government always had been - it was not government by weekly performance-measured outputs, but government by annual inputs.'”

Feasibility

The project was financially and technically feasible. CitiStat uses Microsoft Office programs and software and some other relatively inexpensive software. “For all its success, the CitiStat programme cost the city very little to implement. CitiStat uses basic Microsoft Office programs … as well as GIS mapping software, which costs less than $1,000 … In addition to software, the city hired a small CitiStat staff and renovated City Hall to create the new CitiStat meeting room. In total, the programme cost Baltimore $285,000 to set up10 and carries annual costs of about $400,000, most spent on staff salaries. Needless to say, this investment pales next to the several hundred million dollars it has saved the city.” [11] The Mayor used the city's public works budget to tackle all the expenses incurred for the programme.

The data-driven approach, which had been shown to work in New York's police department, was technically appropriate. “‘The big innovation we stumbled upon was outputs,' he says. ‘We found a better way of governing ... based upon timely and accurate information shared by all, rapid deployment of resources, effective tactics and strategies, and relentless follow-up. It is data-driven decision-making with a collaborative method of questions and answers played out on a common platform." [12]

Management

The staff of CitiStat consists of a director, half-a-dozen analytic staff, and an investigator. It is not a classic management approach, though. “CitiStat is a leadership strategy that permits the mayor and his management team to track, analyse, appraise, diagnose, and improve the results produced by every city agency.” [13]

The staff had good analytical skills and were recruited by the mayor and each of the analysts was assigned specific city agencies to cover. The investigator had the “responsibility for photographing (and re-photographing) egregious past failures, plus new and troubling concerns”. [14]

Bi-weekly meetings were held regularly, which indicates there were mechanisms present to ensure that progress is made. Thus, there were skilled managers in the program who understood the delivery context and were aligned with the objectives.

Measurement

Mayor O'Malley led regular progress meetings to ensure that the project was progressing in the right direction.

The data-driven approach ensured that all the relevant indicators, from crime rates to the mending of potholes in the road, were held in the database and reported on.

The Baltimore administration checked that citizens' requests for service has been satisfied. “It does this by randomly calling each week 100 citizens to see if they are satisfied with the city's work.” [15]

Metrics were used to identify locations where attention should be focused, to ensure maximum efficacy and efficiency. “A good illustration of how data drove decision-making in his mayoralty came after the Clinton administration approved plans to put an additional 200 police officers on Baltimore's streets. The question was where to place them as there were several options available. ‘We could have deployed them equally to all six of the council districts,' recalls O'Malley. ‘Or, if we wanted to be political, we could deploy them to the districts with the highest number of primary voters. Or we could deploy them to crime hotspots in our city, where the vast majority of the shootings, robberies and homicides were taking place.'" [16]

Alignment

There was good coordination between the mayor, CitiStat and the city agencies in carrying out the initiative, for example, in directing police to crime hotspots (see Measurement above).

Bi-weekly meetings were held to maintain coordination between agency and the programme staff. “Managers from each city department then met with the mayor's office every two weeks to answer questions about their results.” [17] The mayor's office uses this data to identify underperformance and press for improvements. When Mayor O'Malley was unable to attend, he appointed deputy Mayor Michael Enright to head the meetings, as mayoral authority was required during the meetings to maintain coordination between different departments and the CitiStat staff.

There was also good alignment between City Hall and Baltimore's citizens, for example, through the use of the 311 number to report problems in the city.

Bibliography

Data to delivery, May 2015, Centre for Public Impact

THE CITISTAT MODEL: How Data-Driven Government Can Increase Efficiency & Effectiveness, Teresita Perez and Reece Rushing, April 2007, Center for American Progress

The Baltimore CitiStat Program: Performance and Accountability, Lenneal J. Henderson, May 2003, The University of Baltimore

What All Mayors Would Like to Know About Baltimore's CitiStat Performance Strategy, Robert D. Behn, 2007, Harvard University

https://martinomalley.com/