Skip to main content
SearchLoginLogin or Signup

Video-recorded Retail CannabisTrades in a Low-risk Marketplace: TradeValue and Temporal Patterns

Published in: Journal of Research in Crime and Delinquency 2018, Vol. 55(1) 103-124

Published onSep 30, 2020
Video-recorded Retail CannabisTrades in a Low-risk Marketplace: TradeValue and Temporal Patterns
·

Abstract

Objectives: This study examines the monetary value of cannabis retail trades and temporal patterns in an open-air market with low legal risks. Method: Video footage detail activities of four sellers and transcriptions were provided by the Copenhagen police. Standard bivariate tests of statistical significance are used to examine the influence of time of day and discretionary activities on trade value and temporal patterns. Results: The average trade was valued at DKK 159 (~US$24) and DKK 121 when excluding 16 outliers. Both the rate and monetary value of trades increase as the day progresses. Twice as many trades are made hourly after sunset and significantly more in relation to national holidays and on days before days with extreme weather. Conclusion: Under the conditions of low legal risks and easy access to supply, cannabis trades are small and tend to increase in value over the course of a day. The temporal patterns in rates follow user availability of discretionary time.

Keywords

drug markets, closed-circuit television footage, cannabis, Christiania

Introduction

Denmark is a small Scandinavian country that has begrudgingly tolerated cannabis retail sales in an area of the capital known as the Freetown of Christiania. Cannabis is still criminalized in statutory law, but a parliamentary compromise in the late 1960s stipulated that the police should not actively pursue retail offenses (Laursen and Jepsen 2002). Visitors to Christiania experience a mixture of self-built houses and colorful characters that is in stark contrast to the neat and orderly surroundings beyond the borders of the Freetown. Social problems and low incomes are overrepresented among the 900 inhabitants of Christiania, and regular police patrols are not undertaken (Midtgaard 2007). The many signs of physical disorder and dilapidation include dogs with no leashes or apparent owners, layer upon layer of graffiti, fires in oil drums, and a general disregard for official rules and regulations. Most notable is the “oppressive feeling” (Amouroux 2009) emanating from cobblestoned Pusher Street, where cannabis has been openly sold for the past 30 years. When cannabis use increased in the 1990s, the legal inconsistency of informal toleration became too glaring for politicians and police, and a large-scale crackdown was implemented in 2004 (Moeller 2009, 2012a).

The cornerstone of the preceding criminal investigation was video surveillance of Pusher Street. This material provides novel insights into the ways in which cannabis buyers behave in a context of low legal risks and easy access to supply. The purpose of this study is to analyze the activities in four sales positions. Standard bivariate tests of statistical significance are used to examine questions on the relation between time of day and monetary value of trades and the association between buyers’ discretionary time and temporal patterns of trades.

Cannabis Markets

Drug markets differ considerably from one another including how users buy their drugs (Pacula et al. 2010). Cannabis is the most commonly used illicit drug, but the market is not well understood because most distribution takes place indoors in social networks that elude direct observation from police and researchers (Caulkins and Pacula 2006; Coomber and Turnbull 2007). While cannabis markets typically cause fewer public order disturbances than other drug markets, they are still important for criminology because users’ spending provides information on the flow of money into the black market (Burns et al. 2013; Kilmer et al. 2013). The best metric for studying this money flow is sampling the number and value of acquisitions directly. However, this is complicated by the network characteristics of distribution and there is a lack of reliable transaction-level data (Bond et al. 2014; Pacula et al. 2010). Without these data, it is difficult to define typical purchases because heavy users account for the majority of trades (Davenport and Caulkins 2016).

Most research on cannabis markets is based on self-reported survey data and interviews. These methods suffer from reliability issues because they typically ask “last acquisition” questions and risk sampling on users and not purchases (Pacula et al. 2010; Sifaneck et al. 2007). Heavy users are undercounted and the latest acquisition may be a “random incidence” (Bond et al. 2014) or simply poorly recalled (Golub and Johnson 2007). Even when these studies sample users at random, they do not sample purchases randomly because relatively few actively purchase their cannabis (Coomber and Turnbull 2007; Davenport and Caulkins 2016). Heavier users account for much of total consumption, and they buy their cannabis. Most of the cannabis that is used was traded for money (Burns et al. 2013; see also Legleye, Ben Lakhdar, and Spilka 2008).1 To better understand how the market looks at an aggregate level, including the monetary value, we need studies that sample on purchases in a way that is independent of the users that make them.

Similar methodological issues are found in research on temporal offending patterns. It is difficult to ascribe exactly when an offense was carried out, especially when the data consist of police records or offender recollections. It is also difficult to discern patterns at an aggregate level because most offense types occur at low density (Ratcliffe 2006; Tompson and Bowers 2012; Kurland, Johnson, and Tilley 2014). It requires a combination of direct and extensive observations to find temporal offending patterns. Prior studies that have examined drug markets and related offenses using video surveillance have analyzed smaller samples to examine microsocial interactions (Piza and Systma 2016; Taniguchi, Ratcliffe, and Taylor 2011). St. Jean (2007) mounted two cameras on a car and drove through the neighborhood under investigation. The video ethnography was used to identify hot spots for different types of crime. These analyses provide documentation that are independent of the object under study and supplement what is known from self-report studies but have not examined temporal patterns over extended time.

At the broadest level, this study examines questions about how much offending occurs and when. Video surveillance of cannabis trades in Christiania collected prior to the crackdown details the time of trades, their monetary value, and hourly rates over an extended period of time. This enables analysis of two types of questions that can contribute to the existing research. The first line of questions concern monetary value and time of day of individual trades, the second line of questions concern temporal patterns of offending, in relation to availability of “discretionary time” (Ratcliffe 2006).

Review of the Relevant Literature

Theoretical Framework

Two bodies of research frame the analyses and interpretation in this study, economically oriented research on drug markets and behaviorally oriented studies of temporal crime patterns. The economic research paradigm is focused on pricing and acquisition sizes (Bond et al. 2014), while the behavioral research examines the social organization and patterns of purchasing (Ratcliffe 2006; Sifaneck et al. 2007). The questions pertaining to temporal patterns are embedded in ecological theories of crime (Brantingham and Brantingham 1995; St. Jean 2007). These lines of inquiry regard offenders’ motivation and the influence of opportunity and are broadly grounded in the rational choice perspective that is commonly used in research on illicit markets (Becker and Murphy 1988; Jacques et al. 2016; Reuter and Kleiman 1986). I combine insights from these paradigms by including discussions about how the structure of distribution affects buyers’ preferences for acquisition sizes (Thompson 2003; Eck 1995).

Cannabis Retail Trades

Most users in the U.S. National Household Survey on Drug Abuse sample (n = 8,339) obtained their latest cannabis from a friend or relative (89 percent) in an indoor setting (87 percent) and did not pay for it (58 percent; Caulkins and Pacula 2006). Davenport and Caulkins (2016) analyzed a later version of the Survey on Drug Use and Health and found that more than half of the most recent acquisitions made by past-month cannabis users were purchases, few of which were from strangers (12 percent) and outdoors (15 percent).

Monetary values of trades vary by intensity of use, but in general cannabis users seem to prefer small acquisitions. Among those who had bought their latest cannabis, 72 percent had acquired less than 10 grams, and a third acquired less than 5 grams. The average monetary value was about US$25 (Caulkins and Pacula 2006). Davenport and Caulkins (2016) found that about half of all purchases made by past-month users were valued at less than US$31 and that 41 percent were valued at US$21 or less. Sifaneck et al. (2007) found that 82 percent of purchases (n = 99) from delivery services in New York City were 3.5 grams or less, corresponding to a monetary value of US$40 to US$50. Golub and Johnson (2004) found that 75 percent of marijuana trades by arrestees in Manhattan (n = 2,979) were for only US$10. In a sample of self-reported purchases (n = 26,611) made by heavier users, Davis and Nichols (2013) found the mean quantity to be much higher at 13.9 grams (SD = 11.1). It is difficult to say how representative these studies are of trades and monetary flows, because there is much variation between users. Infrequent users spent as little as US$5 per day of use while daily users spent US$20 (Johnson and Golub 2007). Other estimates suggest even larger disparities with 1 gram per month for nonregular smokers and 27 grams for daily smokers among 15- to 17-year-olds in France (Legleye et al. 2008).

Trade value also appears to be correlated with the purchase location. Wilkins, Reilly, and Caswell (2005) compared two groups from a household sample in New Zealand (n = 5,800). They found that the people who had bought cannabis from publicly accessible fixed locations known as “tinny houses” (n = 145) smoked more cannabis on their last-use occasion than those who had bought from the “personal market” (n = 342). Tinny house buyers tended to purchase smaller weights, make more trades during a year, and pay a higher price per gram than buyers who used the personal market. Ninety-seven percent of the latest purchases were less than 4 grams and 78 percent were less than 1.5 grams. Jacques and Bernasco (2014) found that the last sale made by 50 street sellers in Amsterdam’s redlight district was no more than 3 grams. Pacula et al. (2010) analyzed a sample of male arrestees (n = 10,370), presumed to be regular users, and found their latest trade size was 10.2 grams (SD = 9.4). Sixty percent of these trades had occurred in private dwellings, where buyers paid lower prices than those made in public exchanges. User-preferred acquisition quantities are affected by the intensity of use and the form of supply, which is broadly referred to as private and public, or networks and markets.

Networks and Markets

Eck (1995) argued that retail-level drug sales can only take two forms: social networks or routine activity marketplaces. These ideal-type distribution forms are differentiated by an open–closed axis where markets are more open and accessible than networks. They also differ on the investments that are required to keep them functioning as economic systems. Participants in social networks engage in interpersonal relationships, and this makes transactions idiosyncratic: person-specific, time-consuming, and inefficient (Thompson 2003; Williamson 1979). The search costs associated with finding a trustworthy seller imply that trades are typically for larger amounts but may be at a favorable price (Pacula et al. 2010; Wilkins et al. 2005).

Conversely, “markets” are characterized by anonymity, accessibility, competition between suppliers, more information on prices, and faster transactions (Williamson 1979). “Drug market” refers to both the national or citywide markets for illegal drugs and the specific marketplaces where concentrated drug sales occur (Reuter and Pollack 2012). Suppliers in open-air illegal markets have more buyers but also face higher law enforcement risks (Eck 1995; Jacques and Bernasco 2014). In this context, coffee shops in Amsterdam (Jacques et al. 2016; MacCoun and Reuter 1997), tinny houses in New Zealand (Wilkins et al. 2005), and Christiania in Copenhagen (Moeller 2012a) resemble markets more than social networks. Cannabis trades in a market structure are less idiosyncratic, and therefore better represent buyers’ preferences. Ease of access to sellers and low legal risks imply that other external factors that affect offending patterns can be studied.

Rate of Trades: Time of Day, Holidays, and Weather

The combination of covert video recordings of geographically concentrated, low-risk and openly accessible cannabis supply provides a unique opportunity to study temporal patterns in buyer behavior. The history of cannabis retail sales in Christiania implies that Pusher Street was a crime attractor. Many users ventured there with the intention and expectation of buying cannabis (Brantingham and Brantingham 1995; Reuter and Pollack 2012; Laursen and Jepsen 2002). Christiania is also a popular tourist attraction in its own right and official estimates range from 1 to 3 million annual visitors. Copenhagen attracts four times as many visitors during the summer months than in winter, but no data are available on seasonal variation in the numbers of visitors to Christiania (Midtgaard 2007).

The available research on temporal offending patterns mostly concerns the effects of darkness on crime rates. Studies have found that a winter peak in robberies is explained by the number of dark hours (Van Koppen and Jansen 1999; Tompson and Bowers 2012). Piza and Systma (2016) noted that 56.5 percent of illicit drug trades occurred after sunset. Theoretically, darkness provides the same discretion that motivates cannabis buyers to conduct their trades indoors and in social networks (Caulkins and Pacula 2006). However, for a consensual crime, such as cannabis acquisition, darkness may rather be an indirect measure of the availability of discretionary time (Ratcliffe 2006).

Social constructions, such as holidays and large events, have also been found to affect and generate offending. Kurland et al. (2014) found that the crime rate was elevated when a large stadium was in use and more potential targets were around. Weather also affects motivation and opportunity, and the relationship between outdoor conditions and crime has been studied for a long time. This research has found that only extremely hot and humid days have a large influence on crime (Van Koppen and Jansen 1999). Tompson and Bowers (2012) found that an increase of one centigrade increased the volume of robberies in London by 1 percent.

Scope of This Study

This study analyzes video-recorded cannabis trades in Christiania from December 2003 to March 2004. The data provide a more reliable expression of buyer preferences than self-report studies because these were collected covertly and are representative of acquisitions rather than users. This means that the analyzes can be used to triangulate the previous research by offering a new and more precise methodological perspective. This novel perspective can add to the conventional knowledge and potentially provide insights into how cannabis markets are currently understood.

The objectives are twofold: first, to generate estimates for the typical value of trades and analyze if there are variations associated with time of day. Next, to extend the analysis of the temporal component and measure variations in hourly rates that relate to buyers’ available discretionary time. This motivation is operationalized in relation to darkness, national holidays, and extreme weather.

Method

Data Source

Copenhagen’s police granted access to the investigative material that had been collected before the crackdown on Christiania. The original video recordings were reviewed and transcribed by police officers and the researcher received physical copies of these transcriptions.

Figure 1. Aerial photo of Christiania’s Pusher Street. The numbered locations are trade positions.

Source: Copenhagen Police (2005).

Three cameras were placed in Christiania, two outside and one inside a building. The two outdoor cameras captured trades in positions 1, 3, and 11 as seen in Figure 1. The indoor camera filmed “position 26” inside building number 26. This position was specifically targeted because undercover police had discerned that it was very active. Each sale position is organized as a small crew, with one or more sellers taking turns and a “runner” transporting money and cannabis to a safe location (Moeller 2012b; see also Johnson and Golub 2007). This system of stationary sale positions implies that the cameras were able to capture detailed images of money and cannabis exchanging hands. The outdoor and indoor positions can be differentiated on the basis of product selection and prices. The outdoor positions typically sold five or six types of cannabis and various joints, but the indoor position only offered one type of cannabis resin (Moeller 2012b). In this position, undercover police had found that the standard offer was “3 grams for DKK 100.” This is a substantial quantity discount compared to what has otherwise been found in international research (Pacula and Lundberg 2014).

Data Coding

The system of stationary sale positions allowed the cameras to capture detailed images of money and cannabis exchanging hands. The cameras were intended to serve different purposes for the prosecution. The data from the outdoor cameras were used to estimate median trade size, and the indoor camera was used to estimate the hourly rate of trades. This is reflected in the ways the units of analysis are coded.

Outdoor.

The outdoor data detail 18 days of trades from three sales positions. The first day of observation was December 2, 2003, and the last day was January 20, 2004. In total there is 89 hours and 15 minutes of footage. The unit of analysis is the individual trade (n = 1,095). Each trade is coded with the hour and minute, and an estimated monetary value, and whether it was resin, joints, or both. The value of each trade was assessed by noting amount of money that exchanged hands. This method is feasible because illicit drug trades typically correspond to rounded monetary amounts (Johnson and Golub 2007; Sifaneck et al. 2007) and the video recordings also enabled a double check with the amount of cannabis or the number of joints. The outdoor data are primarily analyzed for the monetary value and time of day, but some analyses also consider the inclusion of joints. Of the 1,095 trades, 634 were for joints and 20 were for joints and resin. Of these 654 trades, 72 percent included one or two joints and 92 percent was four joints or less. Police analyzed 10 confiscated joints and found they contained 0.33 grams of resin on average (Moeller 2012b), which is identical to quantities estimated in joints in international research (Legley et al. 2008; Ridgeway and Kilmer 2016).

Indoor.

The indoor data describe 89 days of trades at position 26. The first day of observation was December 16, 2003, and the last day was March 15, 2004. In total, there are 1,524.7 hours of footage. The unit of analysis in the indoor data is time intervals (n = 215). Each time interval describes a period where the seller is active on a given date. This implies that the time intervals are different in length and do not begin or end at the same time. Each interval is coded with the hour and minute of opening and closing and the number of trades within this time interval. In total, 62.558 trades are documented in this way.

Measure uncertainty.

This is an administrative data set that is a by-product of policing. It was not originally intended for research purposes, and this entails some methodological weaknesses. I assume that the count of offenses is exact but I recognize that there is an element of uncertainty in the assessment of the amount of money involved. It can be difficult to assess the exact amount of money changing hands when several notes are involved. Video recordings allow a double check to see if the amount of cannabis corresponds to the estimated monetary value, but this is also uncertain because there is considerable variation in prices and in the potency of cannabis (Sifaneck et al. 2007; Pacula and Lundberg 2014; Davenport and Caulkins 2016). A better methodological design would have utilized several coders and an interrater reliability measure. In the present analysis, the magnitude of this measurement error is unknown, but I assume it to be modest because the majority of trades cost DKK 100 or less, and 95 percent cost DKK 250 or less. These amounts were easier to identify because they involved fewer notes, and DKK 50, DKK 100, and DKK 200 notes are quite different in appearance. Larger sums involve more uncertainty, and I have used sensitivity analysis to examine the robustness of the results.

Analytical Strategy

In the first part of the study, the variable of interest is the monetary value of individual trades. The monetary value is in Danish “kroner” DKK, with observations ranging from a minimum 50 to a maximum 6,000. At the time of collecting this data, DKK 100 corresponded to US$15.

The independent variable is time of the trade. To examine variations between different times of day, each trade event was assigned to one of the three temporal intervals: morning (8 a.m. to 10:59 a.m.), noon (11 a.m. to 1:59 p.m.), and afternoon (2 p.m. to 4 p.m.). For analysis, I established the count of trades in each interval and examined the differences between groups with nonparametric bivariate statistical analysis. As the data are for offense counts, a chi-square test (two tailed) could be used, with the categories representing the trifurcated time intervals. The count of number of trades in each interval is compared against a hypothesized even distribution. I also examined if the value of trades changed over the course of a day by evaluating differences in the means in the three time periods with an F test to assess statistical significance. Moreover, I examined the robustness of the findings in a negative binominal regression with the hour as the independent variable. Since most trades are small, there is a large positive skew in the distribution of the outcome variable and the negative binomial regression corrects for this. Finally, I counted that 55 percent of the 1.095 trades were for joints only. The median acquisition was two joints valued at DKK 100. The focus of the analysis is on the monetary value but supplementary analysis considers the role of joints as well.

I conducted various sensitivity analyses to assess how the uncertainty in observing monetary amounts may have affected result. First, I identified 16 outlier trades based on z-scores that have a monetary value of more than DKK 1,100, corresponding to 1.5 percent of all cases. I also examined the results when outliers were defined at a lower threshold of DKK 550. For the supplementary analysis that considers joints, I removed 12 outlier trades that involved more than six joints based on z-scores. Finally, I ran the analyses of trade value and time of day under assumptions of negative and positive observer error. In the negative scenario, DKK 100 was subtracted from all trades in excess of DKK 500. In the positive scenario, DKK 100 was added.

In the second part of the study, using the indoor data, the dependent variable is the trade rate within time intervals. Three analyses were conducted to examine how availability of discretionary time affects trades per hour: time of day, holiday, and extreme weather. The difference in mean rate of trades per hour between the groups was evaluated by a t test.

To compare time of day, I created a cutoff point, “darkness,” defined as 5 p.m., the approximate mean sunset time in the winter. As the observation periods extend across several hours (M = 6.7, SD = 3.6, see Table 1), there are overlaps. The criterion used was that if the observation period overlapped with the cutoff point with more than 25 percent of the observation period duration, then it was removed from the sample. Out of the 215 time periods, 77 were defined as darkness, 99 were defined as “not darkness,” and 39 were removed.

The variable “holiday” was constructed around the national holidays Christmas and New Year’s evening. Christmas holiday in Denmark is December 24 to 26, and the day after New Year’s is an official holiday. To increase the number of holiday observations, I included two days leading up to the events and two days after, reflecting conventional social practices in Denmark surrounding these specific holidays. This resulted in 18 time intervals coded as holiday, corresponding to 8 percent of the sample.

Last, I examined whether extremes in temperature and precipitation have an effect on buyer motivation. Weather data for Copenhagen were drawn from the Danish Meteorological Institute’s (DMI) archives. Days with “extreme weather” were defined by the mean of the high and low temperature for the day being 4 centigrade lower than the mean for the month. Heavy precipitation was defined as more than 10 mm of rain. Forty time intervals were coded as extreme weather, corresponding to 19 percent of the sample. p = .05 was set as the level of significance for all analyses.

Table 1. Descriptive Statistics: Opening Time, Number of Trades, and Trade Frequency.

Descriptive Statistics

Table 1 shows the descriptive statistics for opening time, trades, and rate of trades in the four positions. For position 1, the rate was 13.1 trades per hour over a total of 3,179 minutes or 53 hours. For position 3, it was 6.7 trades per hour over 1,139 minutes. For position 11, it was 15.6 trades per hour over 1,037 minutes. In total, the mean rate is 12.3 trades per hour, weighed to reflect that there are more days of observations from position 1. In the indoor data, there are a total of 62,558 trades in 215 intervals with a mean duration of 400 minutes. This corresponds to an average rate of 41 trades per hour open. Note the substantial variation (SD = 24).

Findings

Trade Value and Time of Day

The average discrete purchase value in all of the 1,095 trades was DKK 159 (SD = 365) with a 95 percent credibility interval from 137 to 180. The large variance is caused by 16 outlier cases with a value of more than DKK 1,100. The mean trade value without these outliers was DKK 121 (n = 1,079, SD = 92). These trades were not evenly distributed across the trifurcated time periods. They mostly occurred in the afternoon, but trades in the morning were not uncommon. Table 2 shows the overall count of trades in the three intervals, the temporal concentration in percentages, mean monetary values, and chi-square and F test values for the distributions.

The mean value of trades increases from DKK 105 in the morning to 119 around noon and then to 129 in the afternoon. There is substantial variation of the means in each period with standard deviations from 86 to 98, but they are statistically significantly different in the F test (p = .026). The pair-wise difference between the mean value of trades in the morning compared to the afternoon is DKK 25 and is statistically significant in the F test (p = .03, 95 percent confidence interval = 2–47). The sensitivity analyses indicated that the results were generally not due to observer error as the mean values do not change much and the difference between the three periods remains statistically significant. The high estimate of the mean, where all observations over DKK 300 are added DKK 100, is 124 (n = 1,076, F = 5.115, p < .01), and the low estimate is DKK 114 (n = 1,078, F = 3.416, p = .033).

Next, I conducted a negative binominal regression to examine the robustness of the correlation between time of day and trade value. With hour as the independent variable and value as the dependent variable, the mean value in the first hour from 9:00 to 9:59 is DKK 94 (with outliers defined as more than DKK 1,100). In the negative binomial regression, the value of trades increases over the observation period (n = 1,079, b = .032, SE = 0.013; p = .014) and the association is also positive and statistically significant when outliers are defined as values above DKK 550 (n = 1,070, b = .025, SE = 0.013, p = .047). This increase in trade value over the day is influenced by an increase in the amount of money buyers spend on joints. The average trade includes DKK 63 (SD = 72) worth of joints, but this is significantly more in the afternoon (M = 72) than in the morning (M = 50).

Temporal Patterns in Rates: Discretionary Time, Holidays, and Extreme Weather

The open access to sellers in Christiania and the low legal risks implied that buyers could acquire cannabis largely at their own convenience. Based on the indoor observation of 215 time intervals, I examined three temporal patterns as inspired by ecological theories of crime (Ratcliffe 2006; Brantingham and Brantingham 1995). The results are presented in Table 3. Rate is measured as the number of trades (T) per hour (H) and the groups are compared in a t test.

Table 2. Distribution and Value of Trades in Morning, Noon, and Afternoon.

First, it was examined whether the rate was different before sunset compared to after sunset. More users have discretionary time in the evening after they leave work and school, and the analysis suggests there were twice as many trades per hour after sunset (M = 54.5, n = 77).

Next, I examined whether the Christmas holiday and New Year’s celebrations had an effect on the intensity of trades. The days surrounding these holidays provide cannabis users with more discretionary time, and the analysis indicates there are about a third more trades per hour, corresponding to 13.5 (SD = 6.2) additional trades, on the days leading up to, during, and following Christmas and New Year’s. This difference was statistically significant at the p = .03 level.

Finally, I investigated whether days with extreme weather influenced buyers’ motivation, measured as difference in rates of trades per hour. There were more trades per hour on days with nonextreme weather, but the difference is not statistically significant in a t test. However, the rate of trades on “the day before extreme weather” was 51.4 (n = 26) compared to the mean for all other days, which was 41.2. (n = 189). This difference of 10.2 trades per hour is statistically significant (t = 2.16; p = .038). The difference is even bigger when the “day before extreme weather” is compared to “days with extreme weather” (n = 39), which have a rate of 37.6 trades per hour (n = 39). This difference of 13.8 trades per hour is statistically significant (t = 2.35, p = .023) as reported in Table 3.

Table 3. t Test for Differences in Mean Trade Rate.

Discussion

This study offers novel data points on illicit drug trades and temporal offending patterns. The analyses are used to triangulate with the previous research and provide a new methodological perspective. The video recordings are independent of the phenomenon under observation, and the sample is a better representation of trades in regard to monetary values than research based on self-reported acquisitions (Bond et al. 2014; Pacula et al. 2010). The data enabled two analyses. The first analysis detailed the monetary values and distributed over the course of a day for individual trades and the second analyses examined how rates of trades per hour vary over an extended time period of 89 days. When comparing these findings with the existing research on cannabis trade value, the implications of the structure of distribution need to be considered. Christiania’s market characteristics of easy access, low risks, anonymous trades, and sellers in competition with each other (Eck 1995; Thompson 2003) imply that the present findings are less affected by the idiosyncratic trade relations in social networks.

The findings suggest that even in Copenhagen’s Christiania district, where cannabis acquisition was a leisurely activity with low risk of legal sanction, people did not buy in uncontrolled ways. On the contrary, the average trade was small. This finding confirms previous studies of selfreported acquisition sizes but contributes to the field by doing so using a new method. The mean monetary value of all trades was DKK 159 and DKK 121 when excluding 16 outliers. Ninety-five percent of trades were valued at DKK 250 or below. In the context of the existing research, this is further evidence that cannabis buyers generally make small acquisitions (Wilkins et al. 2005; Pacula et al. 2010; Caulkins and Pacula 2006; Sifaneck et al. 2007; Golub and Johnson 2004). The 16 outliers corresponded to 1.5 percent of the sample, and this suggests regularity in trade sizes that provides support to the point made by Bond et al. (2014) that the random incidences problem may not be a big methodological issue for the research on illicit drug use.

The video recordings also enabled the collection of a large sample over an extended time period of 89 days. While this sample does not have detailed information on individual trades, the sheer number (n = 62,558) and density of offenses (M = 41 T/H, SD = 24) made possible the examination of temporal patterns at a level of detail that eludes studies that measure the time of offenses indirectly through reported crime (Ratcliffe 2006; Piza and Systma 2016). In sum, the findings of this study demonstrate that there was a high density of offenses, more in relation to days where buyers have more discretionary time. The biggest difference in rates is between day and night. There are almost twice as many trades after sunset as compared to before sunset, but the extra discretionary time associated with national holidays also implied that the rate was higher on these days. Days with extreme weather appear to have indirect impact on the rate trades and more people buy cannabis on days before extreme weather.

Prior research has found that cannabis buyers are sensitive to prices (Pacula and Lundberg 2014; Pacula et al. 2010). Cannabis is a relatively cheap drug, and users have similar sociodemographic characteristics to cigarette users (Davenport and Caulkins 2016). Users should be able to afford larger amounts and realize quantity discounts. The findings that buyers make small trades and half the trades included joints demonstrates a willingness to pay, which indicates that price sensitivity has limited influence on trade sizes in Christiania. The small trade may suggest an alternative behavioral aspect of buyers’ purchasing strategies. Becker and Murphy (1988:678) noted that users of addictive goods will not stock up because individuals “take steps to depreciate the stock more rapidly when it is larger”—that is, the more you have, the more you use. Similar observations are found in marketing research, where it is noted that smokers buy cigarettes by the pack instead of 10-pack cartons even though this means they forego per-unit savings from quantity discounts. Wertenbroch (1998) observed this as a general pattern where consumers of vice goods will voluntarily ration their purchase quantities. He suggested they use it as a self-control strategy by effectively imposing transaction costs on their harmful consumption. Such self-imposed transaction costs will be relatively more important in markets where there is high availability. Buyers in social networks may purchase larger quantities if it is difficult to set up a meeting with a seller and search costs associated with finding competing suppliers are high. Reinarman (2009) noted a similar sentiment that cannabis buyers are concerned with buying high-potency products only if they are worried about access. In other words, a lenient cannabis policy that enables easy access to supply may give users an opportunity to exercise self-control strategies on acquisitions and consumption and may even be an incentive to avoid more potent products.

Next, inspired by ecological theories of crime patterns and especially the notion of discretionary time (Ratcliffe 2006), I examined temporal patterns in trades per hour. The biggest difference was between day and night and this higher rate later in the day may contribute to explaining some of the variation in rates in the three outdoor positions. The sellers displayed variable preference for opening time and had different mean rates. Position 1 generally opened earlier than position 11 and had a lower average rate. Similarly, position 26 opened the latest of all and had the highest rate. However, the mean rate was lowest in position 3, which opened later than position 1, so other factors also exert influence. Each of the three outdoor positions seemed to have different preferences for opening time. Position 1 was especially regular with a range of opening times of only 73 minutes over eight days. If the number of trades a seller made was only affected by users’ availability of discretionary time, then all sellers would open later in the day. Perhaps position 1 catered to niche markets within Christiania where some heavy users make trades early in morning. The overall difference in trade rates between outdoor and indoor position is probably affected by the different time of day of the observations. The difference in rates could also have been affected by the difference in selection of products and prices, which would indicate that buyers prefer lower prices over a wider selection of products.

The increased availability of discretionary time and fewer obligatory activities around the two national holidays in the sampling period are associated with a higher rate of trades. It is a possibility that the high number of tourists in Christiania could have affected these results and also the results on trade value. Tourists are only in the country for a short time, they have plenty of discretionary time, and they have no reason to purchase larger amounts. However, as the data were collected in winter, I assume there are few tourists in the sample. Unfortunately, the proportion tourist among buyers in the different seasons is unknown. It would be interesting to examine seasonal variations to see how summer months and drug tourism affect illicit drug markets. The findings suggest that weather can influence the rate of trades; more people buy cannabis on days before days with poor weather. This association can be hypothesized following the same opportunity logic as discretionary time (Johnson, Taylor, and Ratcliffe 2013). The journey to purchase cannabis feels longer in extreme weather conditions, and buyers may avoid this by going before the forecasted extreme weather. There are higher transaction costs associated with acquiring cannabis outside of discretionary time and on days with extreme weather.

It is generally thought that drug policies exert limited influence on illicit drug use (MacCoun and Reuter 1997; Reinarman 2009), but policy affects how accessible sellers are to potential buyers (Eck 1995; Jacques et al. 2016; St. Jean 2007). Policy has implications for buyers’ purchase strategies, how much they spend per trade (Golub and Johnson 2004; Wouter and Korf 2009). On the continuum of cannabis policy regimes, the Danish approach shares the harm-reducing ambitions and legal ambivalence of the coffee shop system in the Netherlands. The trade-off for a harm-reducing policy is an increase in accessibility and therefore potentially also use (MacCoun and Reuter 1997; Jacques et al. 2016). The findings in the present study indicate that in a regime of low legal risks, users prefer to purchase small amounts and more people tend to buy cannabis when they have more free time. Future research could ask interviewees about their motivations for deciding on trade sizes and purchase timing. How important is access to supply, legal risks, and do self-control strategies aimed at reducing consumption factor in? It would also be of interest to compare the findings on trade value with cannabis retail trades in a market environment with more intensive enforcement. Buyers may choose to make larger acquisitions in order to reduce the number of trades and sanction risk. This would imply more quantity discounts which would affect criminal earnings and throughput capacity and indirectly possibly also increase consumption.

Acknowledgments

The author would like to express gratitude to the editors, anonymous reviewers, and participants in the “Crime Caught on Camera” session at the annual conference of The American Society of Criminology (ASC) in New Orleans, whose comments greatly improved this article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,

authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The author has received travel funding from The Scandinavian Research Council for Criminology.

Note

1. Following Jacques and Wright (2011:731), I refer to transactions that involve money in exchange for cannabis as “trades.”

References

Amouroux, Christa. 2009. “Normalizing Christiania: Project Clean Sweep and the Normalization Plan in Copenhagen.” City & Society 21:108-32.

Becker, Gary S. and Kevin M. Murphy. 1988. “A Theory of Rational Addiction.” The Journal of Political Economy 96:675-700.

Bond, Brittany, Jonathan P. Caulkins, Nick Scott, Beau Kilmer, and Paul Dietze. 2014. “Are Users’ Most Recent Drug Purchases Representative?” Drug and Alcohol Dependence 142:133-38.

Brantingham, Patricia and Paul Brantingham. 1995. “Criminality of Place.” European Journal on Criminal Policy and Research 3:5-26.

Burns, Rachel M., Jonathan P. Caulkins, Susan S. Everingham, and Beau Kilmer. 2013. “Statistics on Cannabis Users Skew Perceptions of Cannabis Use.” Frontiers in Psychiatry 4:1-23.

Caulkins, Jonathan and Rosalie Pacula. 2006. “Marijuana Markets: Inferences from Reports by the Household Population.” Journal of Drug Issues 36:173-200.

Coomber, Ross and Paul Turnbull. 2007. “Arenas of Drug Transactions: Adolescent Cannabis Transactions in England—Social Supply.” Journal of Drug Issues 37: 845-65.

Copenhagen Police. 2005. Virksomhedsberetning [Annual Report]. Copenhagen, Denmark: Københavns Politi.

Davenport, Steven and Jonathan Caulkins. 2016. “Evolution of the United States Marijuana Market in the Decade of Liberalization before Full Legalization.” Journal of Drug Issues 46:411-27.

Davis, Adam and Mark Nichols. 2013. “The Price Elasticity of Marijuana Demand.” Working Paper No. 13-004:1-24, Department of Economics, University of Nevada, Reno.

Eck, John E. 1995. “A General Model of the Geography of Illicit Retail Marketplaces.” Pp. 67-93 in Crime and Place, edited by J. Eck and David Weisburd. New York: Criminal Justice Press.

Golub, Andrew and Bruce Johnson. 2004. “How Much Do Manhattan-arrestees Spend on Drugs?” Drug and Alcohol Dependence 76:235-46.

Jacques, Scott and Wim Bernasco. 2014. “Drug Dealing: Amsterdam’s Red Light District.” Pp. 120-39 in Cognition and Crime: Offender Decision Making and Script Analyses, edited by B. Leclerc and R. Wortley. London, UK: Routledge.

Jacques, Scott, Richard Rosenfeld, Richard Wright, and Frank Gemert. 2016. “Effects of Prohibition and Decriminalization on Drug Market Conflict.” Criminology & Public Policy 15:1-33.

Jacques, Scott and Richard Wright. 2011. “Informal Social Control and Illicit Drug Trade.” Criminology 49:729-65.

Johnson, Bruce D. and Andrew Golub. 2007. “The Potential for Accurately Measuring Behavioral and Economic Dimensions of Consumption, Prices, and Markets for Illegal Drugs.” Drug and Alcohol Dependence 90:16-26.

Johnson, Lallen T., Ralph B. Taylor, and Jerry H. Ratcliffe. 2013. “Need Drugs, Will Travel? The Distances to Crime of Illegal Drug Buyers.” Journal of Criminal Justice 41:178-87.

Kilmer, Beau, Jonathan Caulkins, Gregory Midgette, Linden Dahlkemper, Robert MacCoun, and Pacula Rosalie Liccardo. 2013. Before the Grand Opening: Measuring Washington State’s Marijuana Market in the Last Year before Legalized Commercial Sales. Santa Monica, CA: Rand Corporation.

Kurland, Justin, Shane D. Johnson, and Nick Tilley. 2014. “Offenses around Stadiums: A Natural Experiment on Crime Attraction and Generation.” Journal of Research in Crime and Delinquency 51:5-28.

Laursen, Lau and Jørgen Jepsen. 2002. “Danish Drug Policy—An Ambivalent Balance between Repression and Welfare.” The Annals of the American Academy of Political and Social Science 582:20-36.

Legleye, Stephane, Christian Ben Lakhdar, and Stanislas Spilka. 2008. “Two Ways of Estimating the Euro Value of the Illicit Market for Cannabis in France.” Drug and Alcohol Review 27:466-72.

MacCoun, Robert and Peter Reuter. 1997. “Interpreting Dutch Cannabis Policy: Reasoning by Analogy in the Legalization Debate.” Science 278:47-52.

Midtgaard, Søren. 2007. “‘But Suppose Everyone Did the Same’—The Case of the Danish Utopian Micro-society of Christiania.” Journal of Applied Philosophy 24:299-315.

Moeller, Kim. 2009. “Police Crackdown on Christiania in Copenhagen.” Crime, Law and Social Change 52:337-45.

Moeller, Kim. 2012a. “Estimating Drug Policy Expenditures: Direct Costs of Policing Cannabis in Copenhagen 2000–2009.” Drugs: Education, Prevention and Policy 19:379-86.

Moeller, Kim. 2012b. “Costs and Revenues in Street-level Cannabis Dealing.” Trends in Organized Crime 15:31-46.

Pacula, Rosalie Liccardo, Beau Kilmer, Michael Grossman, and Frank J. Chaloupka. 2010. “Risks and Prices: The Role of User Sanctions in Marijuana Markets.” The BE Journal of Economic Analysis & Policy 10:1-38.

Pacula, Rosalie Liccardo and Russell Lundberg. 2014. “Why Changes in Price Matter When Thinking about Marijuana Policy: A Review of the Literature on the Elasticity of Demand.” Public Health Reviews 35:1-18.

Piza, Eric L. and Victoria A. Sytsma. 2016. “Exploring the Defensive Actions of Drug Sellers in Open-air Markets—A Systematic Social Observation.” Journal of Research in Crime and Delinquency 53:36-65.

Ratcliffe, Jerry H. 2006. “A Temporal Constraint Theory to Explain Opportunity-based Spatial Offending Patterns.” Journal of Research in Crime and Delinquency 43:261-91.

Reinarman, Craig. 2009. “Cannabis Policies and User Practices: Market Separation, Price, Potency, and Accessibility in Amsterdam and San Francisco.” International Journal of Drug Policy 20:28-37.

Reuter, Peter and Mark A. Kleiman. 1986. “Risks and Prices: An Economic Analysis of Drug Enforcement.” Crime and Justice 7:289-340.

Reuter, Peter and Harold A. Pollack. 2012. “Good Markets Make Bad Neighbors.” Criminology & Public Policy 11:211-20.

Ridgeway, Greg and Beau Kilmer. 2016. “Bayesian Inference for the Distribution of Grams of Marijuana in a Joint.” Drug and Alcohol Dependence 165:175-80.

Sifaneck, Stephen J., Geoffrey L. Ream, Bruce D. Johnson, and Eloise Dunlap. 2007. “Retail Marijuana Purchases in Designer and Commercial Markets in New York City: Sales Units, Weights, and Prices per Gram.” Drug and Alcohol Dependence 90:40-51.

St. Jean, and Peter K. B. 2007. Pockets of Crime—Broken Windows, Collective Efficacy, and the Criminal Pont of View. Chicago, IL: University of Chicago Press.

Taniguchi, Travis A., Jerry H. Ratcliffe, and Ralph B. Taylor. 2011. “Gang Set Space, Drug Markets, and Crime around Drug Corners in Camden.” Journal of Research in Crime and Delinquency 48:327-63.

Thompson, Grahame. 2003. Between Hierarchies and Markets: The Logic and Limits of Network Forms of Organization. Oxford, UK: Oxford University Press.

Tompson, Lisa and Kate Bowers. 2012. “A Stab in the Dark? A Research Note on Temporal Patterns of Street Robbery.” Journal of Research in Crime and Delinquency 50:616-31.

Van Koppen, Peter J. and Robert W. J. Jansen. 1999. “The Time to Rob: Variations in Time of Number of Commercial Robberies.” Journal of Research in Crime and Delinquency 36:7-29.

Wertenbroch, Klaus. 1998. “Consumption Self-control by Rationing Purchase Quantities of Virtue and Vice.” Marketing Science 17:317-37.

Wilkins, Chris, James Reilly, and Sally Casswell. 2005. “Cannabis ‘Tinny’ Houses in New Zealand: Implications for the Use and Sale of Cannabis and Other Illicit Drugs in New Zealand.” Addiction 100:971-80.

Williamson, Oliver E. 1979. “Transaction-cost Economics: The Governance of Contractual Relations.” The Journal of Law & Economics 22:233-61.

Wouter, Marie and Dirk Korf. 2009. “Access to Licensed Cannabis Supply and the Separation of Markets Policy in the Netherlands.” Journal of Drug Issues 39:627-51.

Comments
0
comment
No comments here
Why not start the discussion?