Alcohol and homicide in the United States 1934-1995 - or one reason why U.S. rates of violence may be going down.

• Author: Parker, Robert Nash

1. INTRODUCTION

   In the last few years, a great deal of attention has been devoted to the apparent decline in rates of homicide and other kinds of violence in the United States. Commentators debate whether rates of violence are actually declining, and what are the reasons for this apparent decline. The purpose of this paper is to explore the possibility that one reason for the apparent recent decline in homicide may be its relationship to the rate of alcohol consumption during this same time period. As there is a growing body of research that shows a significant relationship between alcohol and violence at different levels of aggregation, in different countries and sub-units of countries, among different types of people, and across time periods, we will also explore the homicide and alcohol relationship by race and by type of alcoholic beverage. There are also the beginnings of a theoretical body of knowledge that would explain why variations in alcohol consumption and availability should be considered part of the explanation for variations in the rate of homicide and other types of violence. These issues will be discussed in detail in this paper, and the results of a new multivariate time series analysis of homicide, alcohol consumption and other indicators for the U.S. between 1934 and 1995 will be presented. The importance of this evidence for violence prevention policy will be discussed as well. Part II proposes that historically, declining homicide rates follow decreases in alcohol consumption. Part III reviews some of the theoretical arguments that have been advanced to explain why alcohol would be a causal factor in homicide and other forms of violence, with references to a number of empirical studies that have found support for this idea. The paper then presents the results of a multivariate time series analysis of the data displayed in Figure 1, with controls for some factors represented in the major theoretical models of homicide in the literature. Finally, the implications of this analysis are discussed in terms of their importance for research on violence and for public policy designed to reduce rates of homicide and other violence in the United States.

   [Figure 1 ILLUSTRATION OMITTED]

2. WHY ARE RATES OF HOMICIDE IN DECLINING IN THE U.S.?

   According to the U.S. Vital Statistics,l the overall rate of homicide in the U.S. has declined steadily in the 1990s. In 1991, the rate of homicides per 100,000 people was 10.5; by 1995, the last year of data available for this report, the overall rate was 8.0.(2) Indeed, media reports indicate that rates for 1996 and 1997 show further declines, with one major media outlet reporting recently that homicide rates are almost as low now as they were in 1970.3 The data given here in Figure 1 confirm this, as the rate of homicide in the U.S. in 1970 was 8.3 per 100,000.

   However, a long term perspective on the issue of whether homicide rates are declining, as shown in Figure 1, might lead to a very different conclusion. Overall homicide rates decreased rather rapidly and more extensively than the decline evidenced in the 1990s during the 1930s and early 1940s, reaching a low of 5.0 in 1944, from a high of 9.7 in 1934.(4) After a brief rise at the end of World War II,(5) the homicide rate declined steadily during the 1950s, reaching a twentieth century low of 4.5 per 100,000 in 1957 and 1958.(6) It is important for understanding recent trends to place those trends in a longer term context as in Figure 1. Is the current downward trend the beginning of a long term decline in homicide rates, as was evidenced between 1945 and 19587 Or is this current trend a short term decline, to be followed by a sharp increase, as was the case between the late 1970s and the early 1990s? A more far-sighted view of homicide rates in the U.S. recommends against the over-interpretation of shorter term trends, and argues against complacency from a prevention policy point of view.

   Another important aspect of the trends in homicide that is not displayed in Figure 1, however, is the fact that hidden by this overall trend in homicide rates is a great deal of variation in those trends themselves, especially when disaggregated by age of the victim. As Blumstein(7) and others have argued, the recent decline in homicide rates, significant as it is, really only applies to the homicides of adults twenty-five years of age and older. The data for youth show the opposite trend, an almost steady increase in homicide rates, especially beginning in the mid 1980s.(8) In addition, as Roncek and Maier have shown, homicide rates vary enormously by city block.(9) Furthermore, Alaniz et al. have shown that rates of assault and other kinds of youth violence vary enormously over short distances within moderate and even small-sized cities,(10) Sherman has also argued that thirty major urban areas in the U.S. account for a substantial majority of all the homicides in the country.(11) Thus, some caution should be exercised in interpreting the declining homicide rate.

   Figure 1 also reveals that for the vast majority of the time covered by these data there is a remarkable correspondence between the rate of alcohol consumption per adult(12) and the homicide rate. Beginning with a decrease in alcohol consumption at the end of World War II, which was followed by a decline in homicide about a year later, a pattern emerged in which changes in alcohol consumption typically foreshadow changes in the homicide rates. The pattern continued for most of the period. For example, alcohol consumption climbed steadily in the middle 1950s while homicide rates began their climb in 1958, peaking in the early 1970s. Between 1958 and 1970, consumption was flat with a slight decline; homicide rates began to decline in the mid-1970s. Again alcohol consumption peaked in the late 1970s, with homicide following by peaking at the end of the decade. By the early 1980s alcohol consumption began a steady decline still continuing in the late 1990s, with homicide again dropping to a low in the middle 1980s. In the late 1980s, fueled by an increase in youth homicide, the two trend lines diverged for the first time in fifty years. By the end of the time displayed here, the decrease in alcohol consumption again foreshadowed a decrease in homicide which began in the early 1990s and continues into the latter part of the decade.

   As was the case with trends in homicide alone, it would be inappropriate to over-interpret the relationship displayed between alcohol and homicide in Figure 1, as will be described later in this article. Alcohol consumption trends differ by beverage type as well; like homicide, alcohol consumption also varies by age. The apparent relationship shown has not received the proper attention from researchers and others asking the question that has motivated this collection of papers, i.e., what explains why the rate of violence seems to be falling in the U.S.? The use of the plural is deliberate, as it would also be inappropriate to claim that a decline in alcohol consumption is the most important or the only reason why homicide rates are falling. A complete explanation of the variation in the rate of homicide certainly involves a number of competing and complimentary explanations. However, Figure 1 clearly raises the notion that alcohol consumption has an influence on homicide to the status of a testable and reasonable hypothesis for consideration. In fact, there is a growing body of scientific research that suggests a theoretically derived and empirically verified causal link between alcohol and violence net of other important relationships, theories, and hypotheses.(13)

3. LINKING ALCOHOL AND VIOLENCE

   A number of studies have reported an association between the homicide rate and alcohol consumption rate. For example, a study of homicide rates in the United States in the early 1980s by Parker found that states with higher rates of alcohol consumption had higher rates of several types of homicide examined.(14) In addition, alcohol consumption interacted with poverty so that places with above-average consumption and above average poverty had even higher rates of homicide.(15)

   A second study estimated the impact of consumption on youth homicide rates.(16) This study examined all U.S. states during the period 1976 through 1983, and found that beer consumption rates were significant predictors of youth homicide rates in five of six age groups by victim/offender relationship homicide rates examined. In both of these studies, additional factors such as poverty and inequality were included as well, and alcohol consumption was found to be a significant predictor of homicide net of these other factors.(17)

   A more comprehensive pooled cross-section time series analysis of youth homicide for states over the period 1973 through 1992 also found a significant net effect for beer consumption on youth homicide rates overall, and specifically for male youth homicide.(18) In this case, controls were included for poverty, inequality, urbanization, beer taxation and other indicators of alcohol policy, and a measure of the impact of the increasing concentration of poor and disenfranchised groups in inner cities.(19)

   Studies outside the U.S. have also reported significant effects of alcohol consumption on rates of homicide. For example, Lenke examined the relationship between alcohol consumption and homicide rates in several European states and found evidence of a significant relationship.(20) Parker examined the impact of alcohol consumption on seventeen European and North American countries in a comprehensive pooled cross-section and time series model for the period 1950 to 1985.(21) Although Parker reports that alcohol consumption did not have any direct effects on either male or female homicide rates, consumption interacted with divorce to increase homicide rates for male victims.(22)

   A second approach to the nature of the relationship between...