Philadelphia Daily News
Thursday, September 25, 1997
Trust us
Mayor's math just does not add up
Rendell's prideful claim of a significant 17 percent drop in crime comes under
fire in light of the city police department's
laborious and confusing record-keeping practices and its apparent disregard
for the
FBI'S rules for crime counting
by Bob Warner and Joseph R. Daughen,
Daily News Staff Writers
The Philadelphia Police Department has been misreporting the number
of city crimes to the state police and the FBI, counting thousands of offenses
that occurred in one year as if they happened in another.
The practice - a violation of the FBI's crime-counting rules - makes a
muddle of the city's reported crime statistics, going back years and probably
decades.
It also punctures Mayor Rendell's recent claim that city crime dropped 17
percent from the first half of 1996 to the first half of 1997.
In fact, what dropped was not the number of crimes reported to police, but
the number of crimes police fed into their computer system.
It's impossible to say what happened to actual crime rates, because
thousands of crimes committed in the first half of 1997 are still waiting to
be added to the Police Department's data.
Rendell's 17 percent pronouncement was intended to defuse criticism that
his Police Department has been unable to reduce violent crime like other
cities had, particularly New York. Since unveiling the figure at a hearing
Sept. 2, Rendell and Police Commissioner Richard Neal have trumpeted the
number repeatedly.
The mayor's arithmetic was based on a claim that there were 53,550 serious
crimes reported in Philadelphia in the first half of 1996. In fact, 23,578 of
those crimes actually occurred in 1995, according to a computer analysis of
Police Department data by the Daily News.
And the 44,330 crimes the mayor reported occurring in the first six months
of 1997 really took place in late 1996 and early 1997, Neal said.
The police commissioner and Rendell spokesman Kevin Feeley insisted that
since the comparisons were based on roughly the same time spans, they were
``consistent'' and thus valid.
Feeley conceded that the mayor didn't know until this week that the figures
he reported on Sept. 2 weren't really for the periods he described.
``His concern is, `Are we getting a consistent picture?' and we are,''
Feeley said. ``There is roughly a two-month lag, so we're getting November to
April as opposed to January to June.
``I think the evidence clearly supports the fact we've tried to be honest
and open and consistent in what we do. There's no question in our minds the
police are trying to be as honest and consistent as possible. Could the system
be improved? Clearly. And it is.''
The Police Department's record-keeping is not nearly so precise. The
numbers Rendell cited from 1996 include crimes committed during all 12 months
of 1995, and a few from 1994 and even before then as well.
So far, the department has refused to provide any of its 1997 data to the
Daily News, so it's not possible to check on Rendell's 1997 numbers.
Neal officially confirmed the Police Department's practice of assigning
crimes to the wrong calendar years after the Daily News discovered that 32,700
crimes committed in 1995 had been counted among the 1996 statistics submitted
to the state police and the FBI.
Spokesmen for the Pennsylvania State Police and the Federal Bureau of
Investigation said they had never been informed of Philadelphia's unusual
counting practices. They also said that practice does not comply with the
FBI's crime-counting rules.
Neal said the department has been following similar practices ``for the
past 30 years that we can determine.'' He said the department needs several
months to investigate reported crimes before the offenses go into the Police
Department's computer system.
``Granted, there may be a delay because of the sheer volume of crimes that
are reported, having it investigated and entered into the system,'' Neal said.
``That's where you get a couple of months of 1995 included in the 1996
numbers, because all the crime has to be reported. We can't just throw it
away.''
The impact on the city's year-to-year crime rates is difficult to assess,
because the volume of miscounted crime has varied from one year to the next.
Using data provided by the Police Department, covering more than 500,000
serious crimes from 1990 into 1996, the Daily News tried to come up with
accurate annual crime counts and compare them to the figures that Philadelphia
had reported to the state police and FBI.
In 1991, the last year of the Goode administration, the number of crimes
reported by the city appeared to exceed the number that actually occurred. In
the four subsequent years, from 1992 through 1995, the official reports appear
to under-count crime, but by relatively small margins, 3.5 percent or less.
Both the commissioner and department spokeswoman Sgt. Susan Slawson said
that other big cities count crimes the same way. But they refused to identify
them or provide any other specific information on practices elsewhere.
Police departments in the nation's four largest cities - New York, Los
Angeles, Chicago and Houston - say they have developed strong record-keeping
practices, providing accurate crime counts with a maximum lag time of less
than two weeks. They denied any reporting practices like Philadelphia's.
In New York City, officials credit their computer system and crime data
with helping them to promptly identify criminal ``hot spots'' and mobilize
police resources to defuse them.
``If you're going to reduce crime you need to have an accurate picture of
the level of crime taking place, right here, right now,'' said Lenny Alcivar,
director of press operations for the New York Police Department. ``Our system
gives us that information right now, not just as something to reflect on six
months or a year down the line.''
Philadelphia's police brass claim they have a separate system for keeping
track of major crimes on a day-to-day basis, providing similar flexibility to
put police where they're needed.
State police Sgt. William E. Krulac, who ran the state's crime-reporting
operation from 1992 until this year, said the state police had been aware that
Philadelphia was struggling with a backlog of crime reports, frequently
delaying the city's monthly reports to Harrisburg.
``What we weren't aware of was that they routinely carry that over from
year to year,'' Krulac said. ``It could put their numbers . . . out of synch
with the rest of the state and the nation.''
The state police collect monthly crime counts from Philadelphia and every
other municipality in the FBI's Uniform Crime Report program. The monthly
reports are due in Harrisburg with about a month's lag time. For example, the
monthly crime counts for September would be due around Nov. 1, according to
Krulac. The state police then prepare their own computer tape to send the
monthly totals to the FBI.
At the end of the year, Krulac said, the state police provide at least two
months for municipalities to submit their December crime counts. ``Our
deadline for the end-of-year UCRs is Feb. 28,'' he said, ``but we routinely
hold the system open into March and sometimes April for Philadelphia and other
municipalities, to give them time to complete their reports.''
Problem is, the Philadelphia Police Department has been running several
months late getting the bulk of city crimes entered into department computers,
according to raw data reviewed by the Daily News.
At some point in the past - when is a mystery - the Philadelphia Police
Department gave up even trying to make its monthly reports conform to the
months that crimes actually occurred. Since the monthly counts were wrong, the
yearly totals were wrong, too.
It appears that murders are still recorded accurately, in the same month
they occur. But it usually takes several months to record rapes, robberies,
assaults, car thefts and other serious crimes. Some crimes don't get into the
police computers for more than a year.
It is this severe time lag that throws in doubt Mayor Rendell's recent
announcement serious crime in Philadelphia had dropped 17 percent between the
first six months of 1996 and the first half of 1997.
Rendell made the statement at a legislative hearing on Sept. 2, barely two
months after the first half of 1997 had ended.
In past years, the department has been unable to come up with accurate
monthly crime counts for at least four or five months, and even then, the
figures have been incomplete. At a meeting yesterday with Daily News
reporters, police officials acknowledged they would not have accurate crime
figures for the first half of 1997 until sometime around the end of the year.
In fact, the 1996 figures unveiled by Rendell on Sept. 2 dovetail almost
exactly with the crime data that police fed into their computers between
January and June 1996.
Officials at the FBI described Philadelphia's practices as a clear
violation of their rules for counting crime.
``There is no provision for contributing agencies to report data from a
prior year in order to develop a full 12-month report,'' said Bennie F.
Brewer, the administrator in charge of the FBI's Uniform Crime Report program.
``It is not UCR policy to accept extrapolations, estimates or any type of
crime report other than actual offense counts.''
Another FBI official, Harlin McEwen, said most major cities face
significant backlogs putting crime data into their computers. But he said he
did not know of any that dealt with the problem the way Philadelphia has by
reporting crimes in months or years other than those when they occurred.
McEwen said he talked personally with Neal this week, after the Daily News
raised questions about Philadelphia's crime-counting, and told Neal that
Philadelphia's procedure was ``incorrect.''
``This is a voluntary program,'' McEwen said. ``We don't have any legal
authority over them on this matter, at all. They have indicated their
willingness to do it right, and they also asked if I could give them
assistance and advice on how to do it better.''
Police officials said yesterday they have already begun improving their
record-keeping with a new computer system that can provide daily counts of
minor crimes. They said it will be six to nine months longer before they can
use it to track murders, rapes and other serious offenses.