The Register-Guard [Eugene, OR], Sunday, June 8, 1997, pp. 1E & 4E

The True Price of Prisons

  • Building more prisons sounds tough on crime, but it's children who pay the hidden costs

    "If we could first know where we are, and whither
    we are tending, we could better judge what to do,
    and how to do it."
    -- Abraham Lincoln, 1858

    By Ed Whitelaw

    When it comes to crime and prisons, we Oregonians don't know "where we are and whither we are tending." This confusion costs us dearly.

    In November 1994, to get tough on crime, Oregonians passed Ballot Measure 11, which requires long, mandatory prison sentences for specified crimes (robbery, assault, homicide, and sex offenses) and treats juveniles who commit these crimes the same as adults. That measure carries a high price tag.

    To any economist, the cost of an action lies not only in the direct outlays for the action but also in the benefits foregone by not taking other actions. The cost of Ballot Measure 11 stems not just from the prisons it forces us to build and operate but also from the opportunities it takes away from us.

    To understand the full cost of Ballot Measure 11, it helps to remember Ballot Measures 5 and 47, which we passed in 1990 and 1996. At the local level, Ballot Measures 5 and 47 cut revenues and pit public safety (police, fire, and emergency medical) against other local services (libraries, parks and health services, for example). At the state level, they shift education funding to the state income tax. Then Ballot Measure 25, which we passed last year, requires a supermajority in the legislature to increase state taxes. Oregonians clearly want to spend less of their incomes on public goods and services. Within this reduced public budget, moreover, Oregonians have imposed clear priorities, notably in how we should deal with violent crime.

    But what perceptions about violent crime have given it such priority in the public's mind and how do those perceptions stack up against measurements of violent crime?

    Consider first our perceptions of crime. In January 1995 and June 1996, the Portland Citizens Crime Commission sponsored surveys of Oregonians' perceptions and attitudes about crime, law enforcement, and incarceration. According to these surveys, large majorities of Oregonians believed "violent crime [had] increased over the [previous] five years" and that "most violent offenders [were] not incarcerated." They were wrong on both counts.

    In Oregon, as elsewhere, the public perception of crime and actual statistics don't match. To understand our measurements of crime, it helps to remember what goes into - and doesn't go into - the statistics we rely on. Most measurements

    Turn to PRISONS, Page 4E

    [A graphic chart at the bottom of the article on p. 1E denotes "Violent offenses per 100,000 population and the Oregon prison population". The horizontal lines stretching from the left margin of the graph rise from 0 to 1,000 prisoners to 2,000 prisoners steadily up to 8,000 prisoners. The vertical lines rising from the bottom of the graph start with 1974 on the far left, then 1975, 1976, and so on to 1995 on the far right. The graph measures two trends. The bottom trendline, representing "Violent offenses," is virtually flat from 1974 through 1995. The number of violent offenses in 1974 is not noted, but the number of violent offenses in 1995 is listed as 523. The top trendline, representing "Prison population," increases more or less steadily (with the exception of a decline of several hundred prisoners from 1987 to 1988) from about 2,000 in 1974 to precisely 7,886 in 1995. The graph is attributed to "Tom Penix/ The Register-Guard. Source: Sourcebook of Criminal Justice Statistics, 1980-95, and the Statistical Abstract of the United States, annually."]
    PRISONS Continued from Page 1E

    of crime - that is, reported offenses and arrests - originate in local police departments. The FBI then compiles the data. When you hear talk about the crime rate, it refers almost invariably to one, several, or all of the FBI's seven so-called index crimes. Four of these seven place the victim in physical danger: murder, rape, aggravated assault, and robbery. We call them violent or personal crimes. The other three are, as Oregon State University economist Arthur O'Sullivan characterizes them, crimes of stealth, not force: burglary, larceny, and car theft. The rest of us usually call them property crimes. These seven categories exclude many other crimes, such as spousal and child abuse (officially called "offenses against family"), other sex offenses, vandalism, simple assault, fraud, drug-related offenses (possession, sale, and distribution of drugs), and drunk driving.

    To put these crime categories in an international perspective, and to illustrate a gap between perception and reality, consider the widely held view that America's crime rates dwarf those of other countries. Citing a 1992 international crime survey from the Netherlands, the January 6, 1996, issue of The Economist stated:

    "You are more likely to be burgled in Australia or New Zealand. You are more likely to be robbed with violence in Spain; you are more likely to be robbed without violence in Spain, Canada, Australia and New Zealand. You are more likely to be raped or indecently assaulted in Canada, Australia or western Germany. America tops the developed-country crime league only in one category: murder. While you are more likely to be burgled in Sydney than in Los Angeles, you are 20 times more likely to be murdered in LA than . . . . . in Sydney."

    But even the United States' world-leading murder rate should be examined closely. Criminologist Richard Moran at Mount Holyoke College recently said, "The first thing you need to know about the murder rate is that it fluctuates." Oregon, for example, reported 4.9 murders per 100,000 population in both 1977 and 1994, and in between reported a murder rate as high as 6.6 (1986) and as low as 3.8 (1990). It's worth noting that Oregon's murder rate in the 1990s has been less than half that of California (rates in Washington State are essentially similar to Oregon's).

    Moran could have said the same thing about the rates of most other crimes anywhere in the country: They go up and down. Consider the FBI's seven index offenses reported in Oregon during 1974-1995. Since the early 1980s, they have gone down three times and up three times. Oregon reported exactly the same number of index offenses per 100,000 population in 1980 and 1995. In 1988, the beginning of the most recent cycle, Oregon reported 7,127 index offenses per 100,000 population. By 1991, the number had fallen to 4,691. These facts belie Oregonians' widely held belief that Oregon has experienced a crime wave in the past two decades. But acting on their erroneous perceptions, Oregonians have expressed special concern about violent crime.

    But consider the number of violent offenses per 100,000 population in Oregon, California, and the United States during 1974-1994. Oregon and Washington have similar violent-crime rates, which are generally less than the U.S. violent-crime rate and far less than that of California. Most important, during the past two decades, Oregon's violent-crime rate has fallen relative to that of Washington, California, and the United States.

    According to the Portland Citizens Crime Commission's January 1995 survey, when they voted for Ballot Measure 11 in 1994, Oregonians believed "that about half of those convicted of violent crimes [were] not incarcerated." They were mistaken. In 1993, according to the Oregon Criminal Justice Commission, we imprisoned over 85 percent of those convicted of murder, manslaughter, Rape 1, Assault 1, Kidnap 1, and Robbery 1. In 1994, the rate had climbed to 93 percent.

    But perceptions govern our behavior and judgment even if they don't reflect reality. And we were so governed in 1994 by the perception that violent crime was exploding and that we weren't incarcerating Oregon's violent criminals.

    With Ballot Measure 11, we turned, like many other states, to more prisons, the most costly option for dealing with criminals. In July 1996, Oregon had 8,509 criminals in its prisons. By July 2006, according to the official (April 1997) forecast, Oregon will have 14,976 people in its prisons. Also according to the official forecast and analysis, about two-thirds of the increase - 4,438 - stems from Ballot Measure 11. At current rates and standards, it costs $80,000 to $100,000 to construct accommodations for each new prison inmate. Splitting the difference at $90,000 puts the price tag of Ballot Measure 11 construction at almost $400 million. The annual direct costs of operation per prisoner, excluding such indirect costs as administration, medical, and court costs, start at $20,000. That would put the direct operating costs at nearly $90 million per year for Ballot Measure 11 prisoners. Prisons cost so much, in part, because they require permanent walls and twenty-four-hour guards.

    Perhaps $400 million up front in construction costs and at least $90 million annual thereafter would be worth it if increased incarceration would reduce the violent-crime rate. Unfortunately, it hasn't done so over the past twenty years. We can perform a quick, simple test that illustrates the bad news. (Alas, we should have performed this test before we considered Ballot Measure 11. It might have given us pause.) From 1974 to 1993, the year before we passed Ballot Measure 11, Oregon's prison population increased 229 percent (from 1,993 to 6,557), while the state's population increased only 34 percent (from 2,266,000 to 3,038,000).

    When making public policy, we shouldn't rely on such a simple test. Fortunately, we don't have to. In their 1995 article, "Estimating the Effect of Increased Incarceration on Crime in California," University of California criminologist Franklin Zimring and his co-authors Gordon Hawkins and Hank Ibser performed a much more rigorous and comprehensive test. They studied California's experience during the 1980s when the state introduced tough-sentencing laws and the state's prison population more than quadrupled (from 23,264 in 1980 to 97,300 in 1990), though the state's population increased only 30 percent. The researchers got the same result, namely, no measurable effect on violent crimes.

    Still, Oregon has spent roughly $1.7 billion on the increase in its prison population since 1974. Granted that's sunk cost (money already spent) and, as any economist knows, should be ignored. But rational taxpayers should ask: What exactly did we buy and could we have bought it for less? During the same period, our violent-crime rate rose and fell several times, ending up, as the graph on Page 1E shows, about where it was in the late 1970s. We bought incapacitation of those we imprisoned, to be sure, and, or course, some punishment. But not much deterrence. That is, we appear to have deterred few, if any, would-be criminals from committing crimes.

    If incapacitation and punishment were the purposes of all that spending, we could have found cheaper, equally effective ways to those ends. When we voted for Ballot Measure 11 in 1994, we already had a pretty good idea of what they were. The Portland Citizens Crime Commission listed some of them in its January 1995 survey: restitution (offenders must pay back the victims of their crime), boot camp (offenders get up early and work hard all day but also can get a high school degree and job training), community service (offenders must do unpaid work such as cut brush or pick up litter), strict probation (offenders see a probation officer twice a week, including unscheduled visits and mandatory drug testing), work centers (offenders are locked up at night and must work at a regular job), and house arrest (offenders must, by electronic monitoring, stay inside their homes). And we liked them, as the commission found: "It would be an understatement to say there is broad, statewide acceptance of sharply increasing the use of alternative sentences or community-based sanctions - there is enthusiasm for the idea."

    Even more valuable than better and cheaper ways to incapacitate and punish criminals are ways of preventing potential criminals from becoming actual criminals. The time is right to find such an approach because, what the January 6, 1996, issue of The Economist calls, "the breeding blip among America's baby boomers" threatens soon to impose on us "30,000 additional young villains" who will spread the peculiarly American virus of viciousness, a violence unique among the industrialized nations of the world.

    Until a 1996 study, "Diverting Children from a Life of Crime: Measuring Costs and Benefits," by a team headed by Peter Greenwood of the RAND Corporation, no one appears to have assessed alternative early intervention methods with compared with incarceration. The researchers evaluated four preventive methods: home visits and day care, parent training, high school graduation incentives, and delinquent supervision. (Many others are examining early intervention strategies, including the UO's Hill Walker and his colleagues at the Institute on Violence and Destructive Behavior.) Their findings run counter to the prevailing policies in Oregon.

    Based on current best estimates of program costs and benefits, investments in some interventions for high-risk youth may be several times more cost-effective in reducing serious crime than long mandatory sentences for repeat offenders.

    Of the four, only home visits and day care (at least as Greenwood's team characterized them) didn't prove cost-effective. The other three did, with high school graduation incentives at the top.

    If we Oregonians were to act on these findings of Zimring and Greenwood, Oregon could decrease its violent-crime rate at lower cost simply by taking the money it's planning to spend implementing Ballot Measure 11 and spending it instead on training parents of very young school-age children, giving cash and other incentives to induce disadvantaged high school students to graduate, and monitoring and supervising youths who have already exhibited delinquent behavior. Not a bad deal. In fact, it may be a much better deal than even this evidence suggests.

    At least as important as what the Greenwood team included in their analysis is what they excluded. They didn't (and don't pretend that they did) account for all the benefits from the preventive methods they examined: long-term reductions in prison costs by diverting youth from lives of crime, for example. What I find especially intriguing is the approach the researchers judged cost-ineffective, namely, "the early home-visit and day-care intervention." Such an approach "works with high-risk youths and their families during the first five years of childhood." The authors then clarify:

    The crime reduction benefits are not reaped until many years after the intervention. However, it affects one form of crime immediately - child abuse [and neglect] by the parents in the targeted families. Studies they cite indicate that certain early childhood services for families would reduce rates of child abuse by about 50 percent. This in turn would yield other benefits - including reduced violent crime - as the children reach their crime-prone years. And there are even more benefits missing from the Greenwood team's calculations - reduced medical and social costs and improved school performance, for example. But here, let's focus on the reduced rates of child abuse cited in that study.

    In all the documents, news reports, political speeches and editorials I've reviewed to prepare this essay, there's not much about child abuse and neglect. In the January 1995 and June 1996 surveys sponsored by the Portland Citizens Crime Commission, for example, the phrase "child abuse and neglect" doesn't appear. The categories of violent crime about which we Oregonians feel so fearful and punitive ignore child abuse and neglect. And Ballot Measures 5, 11, and 47 clearly don't pay much attention to - though they have severe consequences for - child abuse and neglect. No wonder that the 1994 Carnegie Task Force on Meeting the Needs of Young Children refers to American's [sic - Portland NORML] child abuse and neglect as "the quiet crisis:"

    Babies seldom make the news: they do not commit crimes, do drugs, or drop out of school. We don't hear interviews with parents as they anguish over finding decent, affordable child care; we don't notice the unmet prenatal needs of expectant mothers. Policy-makers are rarely forced to contend with these realities. And so, the problems of our youngest children and their parents remain a quiet crisis.

    So let's examine the reality of Oregon's violent crime from yet another perspective. Oregon's 1994 murder rate was 4.9 murders per 100,000 population. Compare that to the fatalities among Oregon's children from abuse and neglect. In 1995, according to data from the Oregon Office for Services to Children and Families and the Oregon Office of Economic Analysis, Oregon had a rate of 12.2 fatalities per 100,000 children less than 5 years old. The graph shows Oregon's 1995 violent-crime rate at 523 offenses per 100,000 population. In 1995, Oregon reported serious child abuse and neglect at a rate of 1,556 incidents per 100,000 children younger than 5 years old - three times the violent-crime rate. This violence against children is staggering - and frightening considering the overwhelming evidence that children who encounter abuse are much more likely to become violent criminals.

    In light of these numbers and facts, how rational is Oregon's policy toward reducing violent crime? Consider only that the Relief Nursery, a successful, private, nonprofit agency in Eugene that is "dedicated to the prevention of child abuse and neglect" estimates its direct and indirect costs of giving its full services to an entire family are $3,000 per year. Recall that the direct costs of imprisoning an adult are $20,000 per year. Could it be more obvious where a healthy and rational society should invest its crime-prevention dollars?

    We have set ourselves up for a very expensive fall. We view criminals as violent and irrational. We've responded with a policy that's also violent and irrational. To restrain and punish criminals, we've chosen the costliest approach. To deter would-be criminals, we've chosen an expensive and ineffective policy. In so doing, we've diverted our increasingly scarce public dollars from less expensive methods of restraint and punishment to prisons and jails. And we've diverted public dollars from education and human services, which could be much more important to reducing Oregon's violent crime in the next decade or two. We have thereby ruined our chances to divert Oregon's at-risk youth away from violent crime.

    We didn't get tough on crime. We got tough on children. On crime, we got dumb.

    Ed Whitelaw is a professor of economics at the University of Oregon. This article appears in the current issue of Oregon Quarterly.



    Back to the All Politics Is Local page.

    This URL: