Forecasting is always risky, but a study that reviews trends in gang problems prior to 1996 should not conclude without considering possible developments subsequent to that period. Very little available research bears directly on future trends in the prevalence of gang localities, but a good deal of attention has been paid to the future of crime in general. Most relevant to gang trends are those studies that deal specifically with violent crime and more specifically with violent youth crime. Although youth gang crime accounts for only a part of violent youth crime, that part is significant, so trends in violent youth crime bear a logical relationship to trends in gang activity.

A major development of the 1990's was the emergence of two conflicting schools of thought on the future of crime. One school contended that serious crime had decreased significantly in recent years and would continue to decrease; the other foresaw substantial, even catastrophic, growth in the next decade and beyond.

Proponents of the "violent crime is decreasing" position based their case on statistics issued by the Federal Bureau of Investigation (FBI) and city police departments showing decreases in arrests for violent crimes in the 1990's. Data in the FBI's Uniform Crime Reports indicated that violent crime decreased by 2.2 percent nationwide between 1990 and 1994. Larger decreases were reported for some of the Nation's largest cities, including New York, Los Angeles, Chicago, Houston, Philadelphia, San Diego, and Detroit. New York reported a 27-percent reduction in major crimes and a 39-percent decline in homicides between 1993 and 1995; Houston reported a 31-percent decrease in major crimes and a 52-percent decrease in homicides between 1991 and 1995. Developments in New York City were particularly widely publicized.

Arrest statistics for 1995 were reported by the New York Times under the headline "New York Sees Sharpest Decline in Violent Crime Rate Since '72: Analysts Begin To Credit New Police Strategies" (Kraus, 1995). The general thrust of this and other accounts was that new police tactics—primarily community policing in conjunction with new technologies and greatly increased community participation in anticrime efforts—were the primary reasons for the decline in violent crime in major cities. In New York, major credit for the success of these policies was claimed by the police commissioner, who predicted "the end of crime as we know it."²⁹ Similar claims were made by police in Boston when 1995 statistics reported a decline in serious crime for the sixth consecutive year. A local criminologist, while agreeing that police policies had contributed to the decline, said that the roots of success ran deeper than police work and credited "a cultural revolution" in attitudes toward violence: "There is a campaign against violence in this country that I've never seen before; we are profoundly concerned with and disturbed by violence, and we're addressing this issue for the first time, and that is making a big difference."³⁰

Those who predicted an increase in violent crime based their case primarily on the traditional concept of the crime-prone age group or age cohort. Most criminologists agree that the most serious violent crime is committed by adolescent and young adult males.³¹ When the size of this group increases, violent crime increases; when it decreases, violent crime decreases. Those who predicted that crime would increase based their predictions on a projected increase in the size of this age group during the next 10 or 15 years. One study predicted that the 14-17 age group would increase 23 percent by 2005; another predicted that the male population ages 15-19 would grow at double the total population rate between 1995 and 2010 (Butterfield, 1996; Duke, 1996).

Some proponents of the "more youth, more crime" position added the concept, also traditional, of the psychopathic or sociopathic killer. This notion was popular in the 1950's and 1970's and provided the basis of a popular novel and movie, A Clockwork Orange (Burgess, 1962). One backer of this position predicted the coming of the "superpredators" (teenage boys who routinely carry guns), who "have absolutely no respect for human life" and who "kill and maim on impulse without any intelligible motive."³² Advocates of this position countered the opposing school by pointing out that while crime in general had been falling, juvenile or youth crime had been rising. One study showed that the number of juveniles arrested for violent crimes increased by 50 percent, to more than 150,000, between 1984 and 1994 and that the number of murders by juveniles had tripled during the same period. Relevant to gang predictions, this study found that the proportion of juveniles who killed in groups rose from 43 percent in 1980 to 55 percent in 1994 (Snyder, Sickmund, and Poe-Yamagata, 1996).

Combining the "more youth, more violence" thesis with the psychopathic killers concept produced a series of predictions of near-apocalyptic proportions. One eminent criminologist predicted "a bloodbath of teenaged violence that will make 1995 look like the good old days."³³ Estimating that teenagers now commit 4,000 homicides per year, the criminologist predicted that this number would increase to 5,000 as the adolescent population grew. A Yale University law professor, citing an "enormous growth in criminal propensities," predicted that "a crime explosion is ready for detonation" (Duke, 1996).

It would appear that accepting the "crime is decreasing" position would support a prediction that youth gang problems would diminish and accepting the "crime is increasing" position would support a prediction of increases in gang problems. The conflicting nature of these two positions and the convincing quality of the supporting evidence for each make the task of predicting gang trends especially difficult.

Although none of the numerical data presented in this Report focuses directly on future growth trends in gang prevalence, some of the data do provide a basis for hypothetical projections. As noted, all projections involve risk and unprovable assumptions, but planning for effective gang policies requires some conception, however speculative, of developments in the gang situation during coming years. The concluding section of this Report will attempt to predict future trends in the growth rate of gang-problem cities.

Predicted Growth Rate of Gang Cities

Data presented in this Report show that there have been very substantial increases in the numbers and populations of gang localities over the past quarter-century. Figure 7 shows growth rates ranging from 2.5 times for all U.S. States to 44 times for South Atlantic cities. What rates can be expected in the future? In contrast to the prediction in the 1982 National Youth Gang Survey report that gang problems would worsen in coming years, the present analysis concludes that the rate of growth in the numbers of gang cities will decrease, possibly accompanied by a reduction in the scope of national youth gang problems.³⁴

This analysis does not attempt to calculate the impact of the many past and present social and law enforcement programs for preventing and controlling youth gang problems or to predict possible changes in the social developments discussed in the sections on explanations. Instead, it uses a set of mathematical calculations based on the notion of a "pool of available gang-free cities," cities that had not reported gang problems by the end of 1995 and were thus in a position to develop gang problems in subsequent years. It assumes that future growth rates are limited by the amount of room to grow afforded by the amount of past growth.

The development of a full and comprehensive research design to test this prediction would be very complex and much too extensive for present purposes. But since effective planning requires some notion of the future scope of gang problems, the concluding section of this Report presents a relatively simple and statistically unsophisticated attempt to support the prediction of slowed growth.

Method Used for Predicting Growth

The basic method used here for predicting future trends is extrapolation from past growth rates to possible future rates. The trend data presented in earlier chapters cannot produce a single trend projection because the data deal with different time periods and different kinds of localities. The analysis will present a limited number of trend calculations, selected on the basis of their predictive potential and relevance to the growth slowdown prediction. Trend projections are based on different kinds of assumptions, ranging from more to less conservative.

The first projection is based on the conservative assumption that future growth will continue on a straight-line basis for the 25 years after 1995 at the same rate as during the previous 25. The data are taken from table 6 and figure 7, which show that the number of gang cities increased 7.4 times in 25 years. Multiplying the number of gang cities in 1995 by 7.4 and assuming that the number of gang cities will also increase by 7.4 times in the 25 years following 1995 produce a predicted total of 2,201 new gang cities by 2000 and 4,442 cities by 2005.

A second prediction uses the same assumption but modifies the projected rate of growth by using data presented by Malcolm Klein (1995), who found 94 gang cities in 1970.³⁵ These calculations produce a figure of 4,131 new gang cities by 2000 and 8,282 by 2005.

A third projection uses the data in figure 6 to make a less conservative prediction based on the assumption that the rate of growth in the number of gang cities that occurred between 1990 and 1995 would continue at the same rate for succeeding 5-year periods. This projection produces a prediction of 4,714 new gang cities by the year 2000 and 9,428 by 2005— figures in the same general range as those in the second prediction.

An additional set of projections is based on the 25-year growth trends calculated for cities in different population categories. These will be useful for comparison with the number of available gang-free cities in different population categories, to be presented shortly. Like the first and second projections, these projections are based on the assumption of straight-line growth for 25 years. For example, for cities with populations between 25,000 and 49,000, 695 new gang cities are projected for the year 2000 and 1,890 for 2005; for cities with populations between 10,000 and 24,000, the projected numbers are 473 and 946; and for cities with populations between 5,000 and 9,000, the projected numbers are 615 and 1,230.

To make the distinctions necessary for this analysis, researchers used the U.S. Department of Commerce, Bureau of the Census' "incorporated place" definition of city units rather than the "units with municipal and township governments" definition, also used by the Bureau of the Census. The latter definition was used in several previous tables (for example, see tables 4, 20, and 21). The reason for using the incorporated place definition is that available census tabulations based on the municipal government definition lump all cities with populations larger than 200,000 into a single category and thus fail to distinguish population subcategories for the Nation's largest cities. This distinction provides an essential component of the slower growth argument—calculations of the numbers and sizes of cities in the four largest population categories.36

The principal body of data used here to support the prediction of a reduced growth rate is a set of calculations on the size of the pool of available gang-free cities. Table 22 displays the numbers and populations of all cities in the United States, the numbers and populations of gang-free cities, and the populations of gang-free cities as a percentage of the all-city population.

One purpose of table 22 is to serve as the basis for table 23, which displays cumulative instead of category-specific figures. The cumulative tabulation makes possible a progressive series of statements on the size of the available pool of gang-free cities, wherein each subsequent statement incorporates the data presented in previous statements. The cumulative figures in table 23 were obtained by adding each successive number to the previous number—the same procedure used in figures 4 and 6.

Note: Data for the 1990's cover the period 1990-95. "Cities" in this table is "incorporated places."

Source: Data on numbers and populations are from Statistical Abstract of the United States, table 45, "Incorporated Places, by Population Size: 1960 to 1990," U.S. Department of Commerce, Bureau of the Census, Economics and Statistics Administration, 1993, p. 24.

Data in each row of table 23 provide the basis for statements on the size of the pool of available gang-free cities in each population category.

Note: "Cities" in this table is "incorporated places."

Source: Data on numbers and populations are from Statistical Abstract of the United States, table 45, "Incorporated Places, by Population Size: 1960 to 1990," U.S. Department of Commerce, Bureau of the Census, Economics and Statistics Administration, 1993, p. 24.

In the four population categories above 100,000, there are 195 cities with a total population of 63.4 million—41.4 percent of the population of all U.S. cities. In each of the four categories, the total number of available gang-free cities is 0—that is, there are no available gang-free cities in any of these categories.

In the "more than 50,000" category, there are 504 cities with a total population of 84.6 million, 55.3 percent of the population of all cities. The number of gang-free cities is 38 with a population of 2.6 million, 1.7 percent of the population of all cities.

In the "more than 25,000" category, there are 1,071 cities with a total population of 104.6 million, 68.4 percent of the population of all cities. The number of gang-free cities is 210 with a population of 8.5 million, 5.5 percent of the population of all cities.

In the "more than 10,000" category, there are 2,361 cities with a total population of 124.9 million, 81.6 percent of all cities. The number of gang-free cities is 1,174, with a population of 23.3 million, 15.1 percent of the all-city population.

In all population categories, there are 19,290 cities with a population of 153.1 million, 100 percent of the population of all cities. The number of gang-free cities is 17,811, with a population of 50.0 million, 32.6 percent of the all-city population. These results are condensed and summarized in table 24.

Analysis of Prediction Data

How do these data bear on the prediction that the rate of increase of gang cities will slow down? Two conclusions are initially evident. First, there is little risk in predicting that rates of increase for cities with populations larger than 100,000, containing 41 percent of the Nation's urban population, will not only slow down but will maintain a 0-percent growth rate. This could change only if future population growth increases the number of cities in the top categories. Even if the year 2000 census adds a few more cities larger than 100,000, however, the number of additional new gang-free cities would be negligible or, more likely, remain at zero. It is almost certain that any city added to the large size categories would already have developed gang problems.

Second, as shown in table 24, as of 1995, cities containing only about one-third of the total urban population were available to become new gang cities. Conversely, cities containing over two-thirds of the urban population had already reached their growth limit, making future growth unlikely or impossible.

Additional evidence with respect to the slower growth prediction is obtained by comparing the growth projections presented earlier with the size of the available pool of gang-free cities—first for cities in selected population categories and second for the total urban population. For cities in the 25,000 to 49,000 category, the predicted number of new gang cities by the year 2000 is 695. Because in the 1990's only 172 cities were available to become new gang cities (table 22), reaching the predicted growth level would be impossible, unless 485 additional cities joined this population category by the year 2000. For cities in the 10,000 to 24,000 category, an increase of 523 new gang cities was projected by 2000 and an increase of 946 by 2005. The number of gang-free cities in this category is 964—18 more cities than the 10-year projection. It is thus theoretically possible for growth in this category to continue at the same rate for a few years after 2005 before exhausting the pool of gang-free cities. However, as discussed below, such continuation is unlikely.

The third and less conservative projection described on page 51 predicts that there will be about 4,700 new gang cities in the United States by the year 2000. However, the number of available gang-free cities in 1995, for all cities except those smaller than 10,000, is only 1,174, about one-quarter of the number of cities needed to meet the prediction. These figures not only support the prediction of slowed growth in the number of localities with gang problems but suggest a future with no growth in the number of gang localities for some population categories.

If, however, cities with populations less than 10,000 are included in the calculations, the number of available gang cities in 1995—about 17,800—is almost twice as large as the number of gang cities predicted for 2005. If growth in the less than 10,000 category were to continue on a straight-line basis after 1995, it would be 2014 before all available gang-free cities were converted to gang cities. The fact that the projected number of new gang cities is substantially larger than the number of available gang-free cities, when all categories except the smallest are taken into account, highlights the role played by cities smaller than 10,000 in the present analysis. Two characteristics of this group of cities are especially relevant.

First, as shown clearly in table 22, the number of cities, towns, and villages in the less than 10,000 category in 1990 was by far the largest of any of the population categories. It included 88 percent of all U.S. cities and exceeded the next-largest category by a factor of 13. Not surprisingly, this category also contained the largest number of gang-free cities. It included 93 percent of all gang-free cities and exceeded the next largest category by a factor of 17. The second characteristic concerns the gang city growth rate in the less than 10,000 category. As shown in figure 19, the growth rate of these small cities and towns between the 1970's and 1990's was substantially greater than that of the larger cities, showing a 22-fold increase compared with a 6-fold increase for cities larger than 10,000. The fact that the under 10,000 category had both the largest number of gang-free cities and the highest growth rate makes it a good candidate for the category with the best growth potential.

The major question then becomes, "Is it possible that continued growth in the number of gang cities among the Nation's smallest cities would be sufficiently likely and sufficiently large to offset a slow-down or stoppage of gang city growth among the larger cities?" Several considerations suggest that no such offset is likely. First, calculations based on population show that cities with populations less than 10,000 constituted only 17 percent of the total urban population, compared with the number-based figure of 88 percent. Gang-free cities made up about half of the population of all cities compared with more than 90 percent on the basis of numbers. It is most unlikely that trends affecting small U.S. cities with a total population of 28 million would outweigh gang-city growth trends affecting larger cities with a total population of 125 million.

A second consideration concerns the size and location of cities with populations smaller than 10,000. The average population of these units in the 1990's was about 1,600. Partial analysis indicates that only about one-third of these communities were located in major metropolitan areas or within 20 miles of larger cities with gang problems.³⁷ The other two-thirds were very small towns and villages, mostly in rural areas, with living conditions characteristic of small towns. This means, first of all, that youth in this kind of community are not subject to the innercity conditions associated with the formation and perpetuation of youth gangs. Second, because of their location, these cities are less susceptible to the influence of larger cities that maintain and sustain a youth gang subculture and less likely to experience migration of outside gangs. Third, the number of youth in the age categories that provide the recruitment pool for gangs is small relative to the numbers of youth in larger cities. Fourth, in most communities of this size, local law enforcement officials are familiar with most gang-age youth, have identified and dealt with the town troublemakers, and are in a position to impede or prevent the formation of local gangs. The degree of secrecy and concealment needed for the formation of local gangs or immigration of outside gangs is hard to maintain in small communities where residents know each other well.

A final consideration concerns the possible effects of research findings on efforts to prevent or inhibit the growth of gang problems in small communities. The finding that smaller towns and villages are more likely to develop gang problems than larger cities could, if widely disseminated, serve to alert local officials to the possibility of an increased risk of gang problems. This could lead to increased vigilance, more active attention to signs of gang emergence, and adoption of procedures designed to prevent local youth groups from adopting the characteristics of criminal gangs. Such activity, if undertaken, could reduce the potential for growth in smalltown gang problems.

Conclusions

The data presented here on prospects for future growth of gang cities may be summarized by two conclusions. First, for all U.S. cities with populations larger than 10,000, the pool of available gang-free cities is smaller than the number of new gang cities projected for the decade following 1995. This suggests that the rate of gang-city growth will decrease and, possibly, for some population categories, will reach zero. Second, the analysis of cities with populations smaller than 10,000, which apparently have the greatest potential for developing gang problems, concludes that the possibility of major growth is limited and that growth rates in these cities would not be sufficient to offset the slowdown predicted for the larger cities. Thus, although the evidence for slowed growth is not conclusive, it provides considerable support for the proposition that an overall increase in the growth rate of gang cities is most unlikely and some support for the proposition that the growth rate will in fact decrease over the next 5 to 10 years.

In conclusion, the prediction of a decreased growth rate is limited to that portion of the gang problem produced by rising growth rates in the number of cities with gang problems. It does not apply directly to other components of the gang problem—principally, the number of gangs, the number of gang members, and the number of gang member crimes. It is theoretically possible, although unlikely, that even if growth in the number of new gang cities were to decrease, the number of gangs, gang members, and gang crimes could increase.

However, if the prediction of a slowdown in the rate of growth of new gang cities proves accurate, the ramifications of this development could affect other components of the gang problem. If the growth of gang cities slows or stops, there could be a concomitant nationwide slowdown in the number of new gangs, gang members, and gang crimes. The slowed growth forecast thus suggests, but does not predict, the possibility of a reduction in the overall scope and seriousness of youth gang problems in the United States.³⁸