Martin O'Malley, the governor of Maryland, thought his question was reasonably straightforward. He was meeting with state parole and probation managers, and O'Malley wanted to know how well -- or badly -- they were monitoring ex-cons in the community. The governor asked how many of the last 50 people charged with murder in Baltimore were under state supervision at the time. The managers were stumped. "They made me repeat the question three times," O'Malley recalls. "It was such an alien concept."
O'Malley's interrogation style may seem alien to some state officials. But it comes as no surprise to anyone who followed his career as mayor of Baltimore, where he created "CitiStat," the intensely statistics-based approach to governance that has since swept city halls across the country. O'Malley is famous for his relentless focus on the data that define government performance day in and year out. Now, he is bringing the same approach to Annapolis, where he calls it "StateStat." The parole and probation office was among the first agencies to run the governor's gauntlet. And its directors weren't particularly happy about it.
As O'Malley tells the story, the flustered parole officials asked how in the world they'd ever find the answer to his question. "I said pick up the phone and call your partners in crime prevention, the Baltimore City police. I'm quite sure they know, and will be able to give you that information sadly and readily." Two weeks later, the officials were back at a StateStat meeting. The numbers they brought with them were startling: 28 of the last 50 people charged with murder in Baltimore had been under the supervision of the state's parole office; 24 of those 28 were in the category of "very intensely supervised."
It was an eye-opening welcome to StateStat for managers who had never seen a governor pay such close attention to what they do. And it's the sort of experience that more and more Maryland officials are growing accustomed to. O'Malley and his staff have rolled out StateStat in virtually every major program area, from transportation to social services, criminal justice and the environment. Indeed, when O'Malley moved from city hall to the statehouse two years ago, the major question wasn't whether he would try to graft the CitiStat model onto state government. That was a foregone conclusion, and O'Malley quickly won legislative approval to give it a try.
The real question was whether the "stat" strategy can work at the state level. The sorts of services cities deliver, from trash pickup to snow removal, are more tangible and more directly manageable than the things state government is responsible for. "In city government, in most instances, there's a pretty short chain of delivery," O'Malley says. "You get a call from a citizen complaining about a pothole. The mayor calls the head of public works, who tells his guys to fill the pothole. So, one, two, three, you're there."
It's different at the state level, O'Malley says: It's harder to force measurable improvements to happen. Take third-grade reading scores. Responsibility for making progress is split among parents, teachers, principals, school superintendents, county school boards, a state educational administration and other players. "It's a much more attenuated chain of delivery," says the governor. Reflecting on the difference between using his statistical approach as a governor, rather than as a mayor, O'Malley puts it this way: "Bigger ship, smaller rudder."
In other words, StateStat poses more political risk to O'Malley than CitiStat ever did. Tracking performance implies some level of responsibility. So an administration that aggressively collects data is setting itself up for criticism if it can't show improvement, too. That's a big reason why only one other state in the country -- Washington State -- has a system in which cabinet officials regularly discuss strategy with the governor's office as the data flow in.
O'Malley believes the risk is worth it. Although it's too early to call StateStat a definite success, the strategy has unquestionably helped Maryland get a grip on a few fundamentals. Overtime in the state's corrections and transit systems is down. The number of assaults in Maryland prisons is declining. There's even been progress on the parole question O'Malley asked about. The number of homicides committed by ex-cons under state supervision is falling.
Most StateStat meetings take place in a dimly lit room one block from the state Capitol. The set-up looks similar to O'Malley's CitiStat meetings -- and a lot like the "CompStat" meetings of many police departments, which is where O'Malley got the idea from in the first place. At tables formed into a squared-off U, the governor's staff sits with the chief and top-level staff of one of the 10 state agencies that come to StateStat each month. A large flat-screen monitor hangs on the far wall -- a scoreboard, in a way. The screen displays the performance measures the agency is expected to live up to, along with graphs and tables showing how it's doing. Matt Gallagher, O'Malley's deputy chief of staff, runs most of the meetings, although the governor tries to make it to every meeting to ask at least a question or two. The interrogations are typically more give-and-take than a grilling, but the expectations are clear: Managers must improve their agencies' performance and if they don't, they'd better have a good explanation why.
That attitude is on full display at StateStat meetings. On a day this summer, the Department of Public Safety and Correctional Services is up for review, and secretary Gary Maynard comes with several of his deputies. Once again, an issue surrounding parole arises: To what extent are local law enforcement agencies acting on arrest warrants for parole violators? Based on the numbers projected on the screen, Baltimore police are clearly doing a good job. Prince George's County, on the other hand, is lagging.
"They're not getting it done," admits one of Maynard's managers. The cooperative tone of the meeting suggests that parole and probation, after the rough start, has gotten on board with StateStat. Department officials and O'Malley's staff spend the next few minutes brainstorming about how to boost Prince George's County's numbers. One idea is enlisting the state police to make arrests in regional blitzes. Another is to begin collecting data on individual warrant teams at the local level, so that StateStat can drill down to see which teams are performing well and which ones aren't.
During the same meeting, numbers are displayed showing the amount of contraband -- drugs, weapons and cell phones, mostly -- confiscated in Maryland correctional facilities. There are clear spikes at discrete times over the past few months, and O'Malley's staff wants to know why. Maynard explains that those spikes coincide with major prison lockdowns and contraband sweeps. He also notes that the prisons have begun using dogs that can sniff out cell phones in addition to drugs. "Can we do more sweeps?" Gallagher asks. Not so easy, one of Maynard's staffers replies: Lockdowns and sweeps take up a lot of staff time and pose safety risks of their own because they seriously disrupt prison routine.
The criminal justice system has been getting a lot of scrutiny in StateStat. That's intentional, says Gallagher, who helped run CitiStat for O'Malley in Baltimore. "From our time in the city, we had a high degree of familiarity with the dysfunction in these agencies." And so it was natural, he says, to focus first on the departments that they knew best and in which they were confident they could make visible progress quickly. Still, StateStat remains focused on what you might call interim measures of performance -- intercepted contraband and crimes committed by parolees -- as opposed to long-term outcome measures such as recidivism rates. "We're not there yet," admits Michael Enright, O'Malley's chief of staff.
Enright, another City Hall transplant from Baltimore, says part of the problem is the matter of jurisdictional control. The state's problems with parole, of course, are highly intertwined with the performance of local law enforcement. But Enright says that's no reason not to measure performance in a broad range of areas, regardless of whether the state might get dinged for failing to make progress in areas beyond its direct control. It's a simple matter of preferring to know what's going on, rather than continuing to operate in the dark.
Enright points to the state's program for providing heating assistance to low-income households. Before StateStat, Maryland would set aside a certain amount of money for the program. Then winter would hit, the fund would get emptied and the program would simply shut down without analyzing who got left out in the cold. "That's insane," says Enright. "I mean, I'll take the heat for shutting the program down when it runs out of money. But we should at least know what the broader need out there is."
O'Malley says he is especially hopeful about how StateStat could energize a lackluster effort that's been going on to restore the health of Chesapeake Bay. Already, Maryland's behavior on bay cleanup is changing, he says, as the state has begun more sophisticated tracking and analysis of where pollution comes from. Now, much more attention is focused on the huge issue of agricultural runoff and on different ways to prevent it -- including the simple step of encouraging farmers to plant cover crops. Meanwhile, the state is using computerized mapping technology to more tightly target its land acquisition policies in ways that protect the bay.
One interesting indication that the effort is making a real difference is the legislature's reaction. Of the $50 million that lawmakers recently appropriated for cleanup costs, not a dime was earmarked. "They left to our discretion to deploy those dollars where they will do the most good," O'Malley says. "And I don't think they'd have done that if not for the discipline and regimen that they saw on display."
Michael Busch, Maryland's Democratic speaker of the House, agrees with O'Malley's assessment. He thinks that StateStat has generally helped the legislature to better see how spending and policy influence state performance. That's especially true in key program areas such as bay cleanup. "If the expectation is that you're going to clean up a certain type of pollution in Chesapeake Bay by doing this and this, and you monitor progress, it allows you to enhance efforts that work and discontinue those that don't," Busch says.
Among state managers and career staff, it's taken some a while to warm up to all the questioning and scrutiny inherent in StateStat. Some, however, embraced it as a chance to make contact with the governor's office with a consistency and intimacy they'd never experienced before. "This is an opportunity once a month to get face time with the governor and his people," says Eric Christensen, a 20-year veteran of Maryland government, who heads up finance and administration at the public transit authority. "It's a way for them to learn our business and the problems we face, and it pushes us to do things that we probably should be doing anyway."
From the employees' side, AFSCME Maryland director Patrick Moran says the union and its members support the general thrust of StateStat and the message it's trying to send Maryland citizens about state performance. On the other hand, the union clearly is chafing a bit at O'Malley's efforts to use StateStat to control overtime costs and boost productivity.
And so the answer to the question of whether the "stat" model can be grafted onto state government is a qualified yes. The O'Malley team has made measurable progress in some key policy areas. What has yet to be demonstrated, however, is the extent to which some of the most crucial outcomes of state government lend themselves to the practice. Still, Maryland presents a promising model that more governors may decide is worth a look.
"The basic principles still apply," O'Malley says of StateStat's evolution from CitiStat. "Human beings work against deadlines. And if goals are well articulated and relentlessly pursued, then smart people will figure out a way to make progress toward them."
These articles are part of a continuing series on public performance measurement focusing on citizen involvement. Support has been provided by the Alfred P. Sloan Foundation. Although the foundation may assist some of the programs described in these articles, it had no control or influence over the editorial content, and no one at Sloan read the material prior to publication. All reporting and editing was done independently by Governing staff.