ECONOMIST Steven Levitt has heard that Secretary of Justice Janet Reno is ''a fan'' of a new academic paper of his. It finds that adding police reduces crime.
That conclusion may sound obvious. But most criminologists find little evidence that hiring more cops in a city reduces crime. In most cases, the number of crimes goes up at the same time as the number of police per capita. That simultaneity could be read that more police cause more crime; or, more likely, that the number of police goes up as crime goes up.
Mr. Levitt's paper, with its contrary view, delights the Clinton administration. It is battling with the Republican Congress to maintain a key provision of the $33 billion anticrime bill Congress passed last summer -- grants to put 100,000 more police officers on the nation's streets by the turn of the century. The House of Representatives has rewritten large chunks of President Clinton's anticrime package. He has threatened to veto any bill that reduces the number of extra police.
''I will do everything I can to protect that,'' Mr. Clinton said earlier this week.
The Republicans say the Clinton bill would pay for far fewer new police than 100,000. But Rahm Emanuel, a White House official bird-dogging the issue, says that in its first four months, the grants have already financed the hiring of 17,000 more cops (at an annual rate) and costs are under budget. ''We are ahead of schedule,'' he says.
If the Clinton measure survives, it will mean a major increase in the number of police in the nation. That number has remained at about 500,000 since 1965.
What the House Republicans have done is reopen the crime legislation, splitting it into six bills. One would provide $10.5 billion over six years for block grants to states and municipalities for crime-prevention measures or hiring police, as the recipients wish. The Democratic measure provides about $6 billion more for these two purposes, but specified the use of the money.
The House Republican bills now go to the Senate. In the House, says Mr. Emanuel, Clinton has the numbers to sustain a veto -- and the Levitt study provides more ammo for a veto.
Sponsored by the National Bureau of Economic Research, Levitt's study found that in 60 of the nation's largest cities, each additional police officer would eliminate eight to 10 serious crimes a year. Using the work of other economists on the costs to society of crime in the way of lost productivity, property loss, medical bills, and quality of life reductions, the value of eliminating those crimes is approximately $90,000 to $100,000 per officer. Given that an additional police officer will receive a salary of about $40,000 and impose nonsalary overhead costs of roughly the same amount, there will be savings for society, Levitt says. The estimates do not include the costs of preventative measures taken by crime victims, lifestyle changes, costs to employers, or legal costs.
These savings, Levitt adds, imply that the current number of officers in the nation is below the optimal level.
Levitt, who has a three-year fellowship at Harvard University, had no idea his paper would prompt a stir among policymakers. What Levitt discovered, after watching the electoral battle for mayor of New York City between David Dinkins and Rudolph Giuliani, is that for political reasons mayors across the nation hire more police before facing election. These police increases, plus changes in crime levels in that year and the following year, gave Levitt a way of estimating the impact of extra police on crime.
Richard Moran, a criminologist at Mount Holyoke College, South Hadley, Mass., hasn't seen the Levitt paper but wonders if the drop in crime numbers at election time actually happens. ''Crime rates are notoriously open to manipulation by police dapartments and mayors' offices,'' he says. Further, he charges both Republicans and Democrats with disregarding all the research of criminologists in devising their anticrime packages.
John Eck, executive director of the Crime Control Institute in Washington, comments: ''Police matter, but not so much in numbers but in what they are doing.'' Research has found that if police are concentrated in crime-prone places and arrest crime-prone individuals, they reduce crime.
Levitt sees no conflict in this research with his findings.