THE EFFECT OF DRUG ON DRUG : EVIDENCE EROM THE MARKETS FOR COCAINE AND HEROIN Jeffrey A. Miron*

Ai?stract—This paper (.•\amiiics the eftect of drug prohibition on the black tries,-^ This approach does not take account of the storage, prices of cocaine and heroin. The paper examines llie raiio of retail to I'armgate for coeaiiic, heroia, and several iegal , and it transportation, distribution, and retailing cosls that exist for compares legal versus black inarket prices for cocaine and heroin. The any product, nor does it recognize that black market sup- results suggest that cocaiue and heroin are substanlially more expensive pliers evade and regulatory costs typically incurred by than they would be in a legalized market, but to a lesser degree than siiggesled in previous research. legal suppliers. Thus, the difference between the farmgate price and the retail price overstates the effect of prohibition. I. Introduction This paper reconsiders the relation between drug prohi- bition and drug prices. The results are qualitatively similar HB effect of drug prohibition on drug prices is an to those in the previous literature, but they suggest a less T important question in the debate over legalization ver- dramatic effect of prohibition. And by allowing for various sus prohibition. Existing analyses suggest that prohibition factors that determine drug prices, the results provide more has raised prices dramatically, making drugs len, twenty, or convincing evidence of these effects. even hundreds of times more expensive than they would be Section II presents data on the ratio of retail to farmgate if legal.' From both theoretical and empirical perspectives, price for cocaine, heroin, and legal goods such as chocolate, however, the standard conclusion is potentially too strong. coffee, tea, beer, and tobacco. The data show that retail On the theoretical side, the presumption that prohibition cocaine and heroin prices are hundreds of times the costs of raises drug prices implicitly compares prohibition with the raw materials used to make these goods, consistent with laissez-faire. The relevant comparison, however, is between earlier analyses. The data also show, however, that the retail prohibition and the taxation-cum-regulation regime that price is sometimes hundreds of times the farmgate price for would apply if drugs were legal. Black market suppliers legal goods as well. Thus, the high value of this ratio for face low marginal cosls of evading tax and regulatory cocaine and heroin does not by itself indicate a substantial policies, in that they normally hide Iheir activities from law effect of prohibition. enforcement authorities; thus, they enjoy a cost advantage thai partially offsets the increased costs created by prohibi- To infer the effect of prohibition from these data, one tion. In addition, prohibition differs from a taxation-cum- must decide which legal good provides an appropriate regulation regime in the amount of enforcement, the nature model for a legalized cocaine or heroin market. If the of compliance incentives, the degree of market power, and appropriate benchmark is relatively unprocessed goods, the level of advertising. These differences can weaken or such as coffee beans in a grocery store, the data suggest that even reverse the standard presumption (Miron, 2001). black market cocaine and heroin cost perhaps hundreds of On the empirical side, most analyses of prohibition and times their legalized prices. If the appropriate benchmark is drug prices simply note that the raw materials used to more processed goods such as espresso at Starbucks, then produce drugs sell at low prices in producer countries while the data suggest black market cocaine is 2 times the legal the finished products sell at high prices in consumer coun- price and black market heroin is 6 times the legal price. The right nu)del is likely between these extremes, but 1 suggest

Received for publication May 21. 2001. Revision accepted lor publica- that the espresso benchmark is not implausible. tion June 28. 2002. Section III examines the prices of legal cocaine, mor- "^ Depariment of , University. I am grateful to the Barhart Foundation for tinaneial support and to phine, and heroin. Cocaine is currently prescribed as a Angela Dills ibr research assistance. I diank Elena Alvarez, Ernie Berndt, topical anesthetic, and morphine is widely used as an Shane Hunt. Mark Kleiman, Aviv Nevo, Rieardo Rocha, Fiona Scott- analgesic. In addition, cocaine, morphine, and heroin are Morlon, Riehard Spady, Juan Tecco, Norbert Telnion, and Francisco Thoumi for help in finding and understanding data, Carolyn Travers and used legally for scientitic, analytic, and research purposes. Robert Janice of the DEA for help in obtaining and understanding the Quality considerations aside, black market cocaine has STRIfJE data, and William Rho

The Review of Economics ami Siatistics, August 2003, K3(3): 522-5M) ir: 2003 hv lhe President ;iiid fellows ol Harvard College and the Mas.sachusetis Institute olTechnology i:)RliC. PRICES 523

Scclions II and HI ihus suggest thai currcni cocaine and t(ik Ci.H.-x l.i.'\r .'XNti BLACK MARKI-;I OK heroin prices are suhslanTially higher than they would be in Price of AniounI a legal market, bii( io a lesser degree than suggested in Required, per Gram earlier work. One reason the difference is that earlier Produci Date 1.(11. alion Market of Pure CtlCL analyses assume prohibition causes (he entire increase in Coea leaf ' IMK Peru KirnigaiL' $0.36-$0.57 priee from I'aF'mgate to retail. In addilion. earliei' analyses Cocaine I'm U.S. Relail $122.00 fail tt.) note ihal although prohibilion imposes costs on blaek RalH) of reiail t(i farmgale price: 262 market suppliers, it facilitates evasion of costs typically borne by legal suppliers. added. A white substance forms al the bottom. This Section IV tlierelore examines the costs incurred by legal white substance, when dried, is coca paste. suppliers but evaded or avoided by black market suppliers. 5. The coca paste is crumbled and diluted in acetone; a These include, for e.\aniple, ; environmental, safety, liuid containing the alkaloid is then decanted. Hot air and health regulation; labor market regulation; and adver- is added to evaporate the acetone, leaving the alkaloid tising expenditures. The results suggest that such costs condensate. This substance is pressed to remove water constitute about 5()'>(^ t)f the priee of legal products. and resins; the remaining subslance is cocaine base. Section V then examines the costs imposed by prohibition 6. The cocaine base is diluted in acetone and mixed with enforcement. The results suggest that prohibition's effects acetone, ether, and hydrochloric acid. The ether and on labor, eapital. and materials costs are of aboul the same acetone arc decanted to leave the salt formed from magnitude as the tax and regulatory costs avoided by black cocaine base and hydrochloric acid {cocaine hydro- market suppliers. The full effect of prohibition is larger if ehloride). After drying, this substance is powdered prohibition affects factor proportions or economies of scale. cocaine. But the elteets on factor prices are consistent with relatively modest overall effects. Table I presents data on the price of coca leaf and Section VI concludes. cocaine.' The first row displays the price in Peru of the amount of coca leaf neeessary to produce a pure gram of II. Thf Prices of Cocaine, Heroin, and Some cocaine hydrochloride (CHCL). The second row displays Products the retail price per pure gram of CHCL in the United Stales. The ratio of retail to farmgate price is 262. This section examines the effect of prohibition on drug Heroin is a semisynthetic compound derived from the prices by estimating the ratio of retail priee to fumigate opium poppy. Fapaver sotnnifentm. Producing heroin from priee for cocaine, heroin, and several legal gotKls. Prohibi- opium consists of the following steps (Krivanek, 1988, p. tion is not the only reason this ratio might differ aeross 106): eommodities; production, storage, transportation, distribu- tion, and retailing costs all contribute to this ratio. But if the 1. Incisions are made on the immature seed capsule. A ratios for cocaine and heroin are consistently larger than milky exudate emerges; the dried exudatc is brown, those lor legal goods, this suggests an important role for raw tipium, which is marketed by the opium farmers. prohibition in raising drug prices. This ratio is also the main 2. Raw, brown opium is soaked, heated, and (litered; thi.s indicator of prohibition's effect employed in previous work. produces a brown powder, which is morphine base, or No. I heroin. It takes about 10 kg of opium to make A. llic Piodiiclion and Pricini; of Cocninv and Heroin 1 kg of morphine base, whieh in turn produces about 1.1 kg of heroin (Lewis. 1984). Cocaine is an alkaloid of the coca plant. Extracting 3. Acetylalion of the morphine base produces a powdery cocaine from coca involves the following steps (Morales, gray substance known as No. 2 heroin or heroin base. 1989. pp. 76-86); This is not water-soluble and not injeetable. 4. No. 2 heroin can be refined into No. 3, or smoking, !. Raw coca leaves are picked by hand, dried in the sun. heroin, a granular, soluble salt that can be gray, and marketed. brown, bluish, or pink. This common form of heroin 2. Dried coca leaves are mashed with water and sulluric has a morphine content of about 30%. acid; this produces a brownish, acidic liquid that 5. No. 3 heroin can be refined further to give No. 4 contains the cocaine alkaloid. heroin, a white powder. It ean be up to 90% tnorphine. 3. The aeitlic liquid is introduced into a new pit. where the acid is neutralized by adding a ba.se. An organic Table 2 presents data on the price of opium and heroin. solvent is added, after which a fluid that does not The first seven rows display the prices in various countries contain the solvent is decanted. 4. The second decanted fluid is added to a eontainer of • The dala sources lor ail latilcs are yivon in an appendix itiai is available water and sulfuric acid; sodium carbonate is then upon requcsi. This information is aiso avaiiatite in Miron (2001). 524 THH REVIEW OF ECONOMICS AND STATISTICS of the amount of opium necessary to produce one gram of Aid [- }. —\ CHIK (11 heroin. The final rovv^ gives the retail price of heroin in the Priee per . Assuming the amount of opium neces.sary to F'oiind of Froduei flale l.oeation .Vlarkei Cocoa Beans produce one gram of heroin eosts $1.00. Ihe data imply a retail-to-farmgate ratio of 844. Coeoa beans Ql. !W'J (Ane d'lvoire 1-arm gale $0.30 Cocoa beans 1 998 Loiidoii/W^' tix/im $0.76 Milk ehoeolaie 06/99 Boslon Retail $26.44 B. The Production and Pricing of Legal Goods Cocoa powder, regalav 08/99 Bosioii Relail $9.40 Coeoa powxier. dulcli 08/99 tHoslon Relail $18.83 I now examine the produetion and pricing of several legal (."Lip of liol elioeolate O8/9M Bosion r

TABI.F 2.—PKICF.S OPIUM ANI"* BLALK M..M

TARI I; 4,—pRicFs FOK COFFF.F TABLK 6,—PRICIIS IOR BARI.I.Y AND BF.FR Price per Ib price per kg Producl Date t.ocation Market Coffee Product Date Location Market of Barley

Raw beans [arahica) Q4, 1998 Colombia Farmgate $0.75 Malting barley 04/99 U,.S, I'anngate $0,096 Raw beans {anihicu) Q4, 1998 Brazil Farmgate .$0,88 Case of beer, liquor store 08/99 Boston Retail $I3.32-SI7.78 Roasted ground beans OS/99 Boston Retail $2,60 Bottle of beer, bar or Roasted whole beans 08/99 Boston Retail $6,36 restaurant 08/99 Boston Retail S5^ 13 Cup of coffee 06/99 Boston Retail $25,43 Espresso drink 06/99 Boston Retail $111,30 Ratio of Retail to Product 1-arm gate Price Ratio of Retail to Product Farmgate Price Beer in liquor store 139-185 Beer in bar <"ir restau .'i56 Roasted ground beans 3 3,^ Roasted whole beans 7-8 Cup of coffee 29-34 Hsprcs'^o drink 126-148 retail-to-farmgate ratios exist for legal goods, including those with distribution and retailing pattems that are similar to what would plausibly exist for drugs in a legal market, at least raises C. Discussion the possibility that prohibition's effect is far smaller than The critical question in drawing conclusions from these indicated in previous research. data is which legal good provides a plausible model for Independent of which comparison provides the right legalized cocaine or heroin. One view is that the legalized benchmark, there are a number of possible biases in the market would resemble, say. ground coffee at the grocery calculations reported above, and these suggest smaller store. Under this view, the data suggest cocaine sells at 80 rather than larger effects of prohibition. First, more process- times its legal price, and heroin at hundreds of times its legal ing takes place before the first market transaction in the price, A different view is that the legalized market would legal products than in the cases of cocaine and heroin, resemble, say, the market for espresso at Starbucks. Under implying that the retail-to-farmgate ratios are understated this view, the data suggest cocaine sells at only about twice for legal goods. Second, the prices of legal goods in spe- its legal price, and heroin at about six times its legal price. cialized locations, such as highway rest stops or mi^vie In a legal market, drugs would presumably be available theaters, are typically higher than ihose considered here. both in relatively unprocessed bulk quantities at low prices Third, the data source used to obtain price data likely (e.g., bulk CHCL at drug stores) and in relatively processed overstate transaction prices faced by actua! drug consumers small quantities at high prices (e.g., crack, smoking heroin, as opposed to DEA agents (Miron, 2001, pp. 75-76),^ coca tea, and opium drinks at bars and coffee shops). Thus, A different issue is that the comparisons here implicitly as with many legal goods, low and high prices would exist assume legal drugs would be taxed and regulated like any simultaneously for different versions of the commodity. other good. In practice, drugs would likely be subject to The transactions used to compute the price of black market higher than usual taxation. Excessively high taxes would C(K:aine and heroin, however, are for small quantities in rela- result in a black market, but experience with and tively processed form (e.g., small amounts of crack). Thus, the cigarettes suggests taxes can raise prices by a factor of 2-3 more appropriate comparison is plausibly with the prices of the without generating a black market. relatively processed legal products, such as espresso, implying The bottom line is therefore that previous analyses have effects of prohibition at the lower end of the range discussed overstated the effect of prohibition on prices, and under plau- above. This conclusion is not iron-clad; the purchase of an sible assumptions this overstatement has been substantial. espresso at Starbucks differs in many respects from purchase of cocaine at a crack house. But the fact that substantial III. The Price of Legal Cocaine, Morphine, and Heroin A different way to determine the effect of prohibition on TAB[,H 5,—PRICIZS FOR TF:A prices is to compare black market and legal transactions.''' Price per Cocaine is used legally in the United States as a topical Product Date Location Market !b of Tea Dried tea 1999 Sri Lanka Auction $0.50-L{)0 •^ A possible bias in tbe other direction is iliai cocaine and heroin Box of tea bags, regular 07/99 Boston Retail $5,98 consumers might purcha.ses small amoimts at high pcr-tinit prices because Box of tea bags, specialty 08/99 Boston Retail $25,69 Ihey are liquidity-constrained. An alternative explanation for such behav- Cup of tea 08/99 Boston Retail $174.83 ior is that consumers are committing to a low consumption rate, '• Moore (1990) and Morgan (1991) compare the price of legal and black Priee Ratio. Product market cocaine. The legal price quoted by Moore is substantially below Retail/Auetion the current legal price, but it is not based on a transaction in bulk cocaitie. Box of tea bags, reuular 8 The price cited by Morgan is also substantially below the current legal Box of tea bags, specialty 34 price, but it is a price available to ''investigators.'" which may not be the Cup of tea 233 general market price. Nevertheless, there appears to have heen a substan- tial increase in the price of legai cocaine over ihe past decade. 526 THE REVIEW OE ECONOMICS AND STATISTICS

TABLE 7.—PRICLS TOR TOHACCO TABLE 8.—LEGAL VERStJS BLAC"K MARKET PRICES KJR C(K'AINE Price per pound Amount Price Unii pncc'' Product Date Locaiion Markei of(obaccQ Market (S)

Cured tobacco 1998 U.S. Far tn gate $ 1.83 Prescripuon, MaMinckrodt 5 324.19 ()4.K4 Cigarettes 06/99 Boston Retail $55.23 25 1710.94 68.44

Ratio of retail to fjirmgate price. lobacco; ?•() Prescription. A-A Spectrutn 5 335.00 67.00 25 1650.00 66.00

Research, Company 1 1 83.55 Research, Company 2 10 858. f.0 85.60 anesthetic, mainly in the upper respiratory tract (Catterall Analytical, Company 3 100 49.30 and Mackie, 1996, p. 338). In addition, cocaine is used Analytical, Company 4 1 112.25 1 12.25 legally for certain scientific, analytic, and research purposes. Black market 1 106.82 For example, small amounts of cocaine are used to test 5 84.34 blood, urine, and hair samples for the presence of cocaine 10 69.98 metabolites. Heroin is not used legally as medicine in the 25 46.79 100 34.29 United States, but it is used for scientific, research, and ' Per unil (if pure i'ix:auic. analytic purposes.^ prices of cocaine in the United States and in Europe, where A. Prices for Cocaine at least two manufacturers exist, and concluded thai "prices Table 8 provides data on the price of legal cocaine. The in the foreign markets are between thirteen and twenty-two first four rows are from the Red Book, a standard catalogue percent of the domestic price for a kilogram of cocaine" of wholesale pharmaceutical prices used by pharmacists, (Federal Register, 1998, p. 38). hospital dispensaries, HMOs, and the like. These are prices The estimate of the monopoly markup provided by the at which pharmaceutical manufacturers are willing to sell DEA investigation was based on transactions of at least their products. The next four rows report the prices at which 100 g. Dividing the 100 g price from Company 3 by 5.7, [he various companies buy or sell cocaine used in reseai'ch, midpoint of the DEA estimates of the markup, implies the scientific, and analytical products. The last five rows give black market price of cocaine is four times the legal (com- black market prices for various transactions sizes. petitive) price. The data show that black market prices for cocaine are similar to the legal prices; indeed, the legal price exceeds B. Prices for Legal Morphine and Heroin the black market price for some transaction sizes.'^ This Table 9 presents data on the legal and black market prices comparison is potentially misleading, however, because the of morphine and heroin. The first seven lines give the price legal prices probably include a monopoly markup. There is of prescription morphine from the Red Book. The price per currently only one company that legally imports substantial gram ranges from $3.50 to $10.61. The next three lines give amounts of coca leaf into the United States. This company the price of morphine available for scientific and research extracts the cocaine in the form of paste and sells it to a purposes from the company Sigma-Aldrich. The price per second company. This second company refines the paste into bulk cocaine and sells it to other companies. The second company is thus the only legal manufacturer of bulk TA!![..F 9.—LEGAL vERstis BLACK MARKET PRICHS FOR AND HL;ROIN cocaine in the United States, which suggests its prices contain a monopoly markup. Price Unit Price" Markei Amotint ($) A recent legal case provides information on this markup. Morphine, lied Book. A-A Spectrtim 28.50 5.70 In 1995, a company that purchases bulk cocaine from the 95.00 3.80 U.S. manufacturer sought permission from the DEA to lOOg 350.00 .1.50 import from a European manufacturer selling at a lower Morphine. ReJ Book. Mallinckrodt 5g 53.05 10.61 price. Under current iaw, the DEA must allow importation if 25 E 265.25 10.61 "competition among domestic manufacturers is inadequate" 50 g 530.05 10.60 lOOg 1,061.00 10.61 (Federal Register, 1998, p. 28). The DEA compared the Morphine. Sigma-Aldrich 25 mg 20.00 8(H}.00 1 g 155.00 l55.Otl ^Cocaine and morphine are Schedule II drugs under the Conirolled H 311.00 62.20 Substances Act of 1970. This means they have accepted medical uses and can be legally prescribed under certain conditions. Heroin is a Schedule I Heroin, Sigma-Aldrich 25 mg 117.(XI 4.680.00 drug, which means it has no currenlly accepted medical use in the United 25 mg 1,444 Stales. Heroin, Black Markei i S 484 "The comparisons between black market and pharmaceutical cocaine 275 understate the price of pharmaceutical cocaine by abotit 11% because 100 g 113 these are prices for cocaine hydrochloride, which is 89% cocaine by weight (Grinspoon and Bakalar. 1976. p. 74). DRUG PRICES 527 gram, for a different set of iransaction sizes, ranges from In 1996 approximately 4 million persons in the United $62.20 to $800.00. The next line gives the price of heroin States used cocaine, some of them many times, (U.S. available from Sigma-Aldrich for scientific and analytical Department of Justice. 1998, Table 3.S3, p. 245), and (here purposes. The price per gram for a 25 mg transaction is were 144.180 emergency-room episodes related to cocaine $4,680. The last four lines give the price per pure gram of (U.S. Department of .lusticc. 1998. p. 253). Thus. Ihere were black market heroin for different transaction si/cs. aboul 3.6 emergency-room episodes per hundred persons The first issue in interpreting this information is that the who used cocaine. Of these 144.180 episodes, however, priee per pure gram of morphine from Sigma-Aldrich i.s only 357f were caused by overdoses or unexpected reaction, markedly higher than that from the two Red Book sources. as opposed to chronic effect, seeking detoxification, or Holding transaction size constant, the price from Sigma- withdrawal, and only episodes in the first two categories can Aldrich i.s 5.9-10.9 times higher. reasonably he attributed to poor quality.'" Further, overdose I assume that this premium reflects additional costs, such and unexpected reaclion can occur even when ihe quality is as guarantees about purity, related to the specific uses of the known: many persons consume high doses o\ alcohol, for Sigma-Aldrich products. Assuming pharmaceutical heroin example. Thus, ihe number of emergency-room episodes would be cheaper than the re.search-quality heroin by a attributable to quality concerns is modest. similar factor, this suggests using 4680/10.9, or S429.36, as A second caveat is that drugs might be more available in the price per pure gram of heroin in legal transactions of a legalized market, suggesting ihc full price would drop roughly 25 mg. more than the monetary price. Fxisting evidence, however, The remaining issue is whether, as with cocaine, the legal suggests drugs are already widely available. Over the 1985- prices reflect a monopoly markup. This is possible, but there 1997 period. Ihe percentage of high school seniors stating it are currently two companies that legally import opium and was "lairly easy" or "very easy" lo get cocaine (heroin) derivates into the United States (Stecklow and Karp. 2000). always exceeded 40'7< and frequently exceeded 50% (20% Thus, the markup is plausibly smaller than for cocaine, but or 30%) (U.S. Department of Justice. 1998. p. 162). These to err on the conservative side I assume it equals the cocaine responses .suggest a fair degree of availability, since over the markup of 5.7. Combining this with the assumption that Ihe same period the percentage of high school seniors admitting price of 25 mg of heroin from Sigma-Aldrich should be to having used cocaine (heroin) in (he previous twelve deflated by a factor of 10.9 and cotnparing this with the monlhs ranged belween 3.l'7r and 13.1% (0.4"7r and 1.2%) price per pure gram for black market transactions that arc (U.S. Department of Justice, 1998. p. 237). roughly 25 mg, we find that black market heroin is roughly 19 times the legal price. An alternative approach to estimating the legal price of IV. Ihe Costs of Producing Legal heroin is to combine information on the price of legal morphine with information on the comparative potency of The results presented above stiggcsl a more modcsl effecl heroin and morphine. As shown in Table 9, morphine is of drug prohibition on drug prices ihan presented in previ- available from legal manufacturers for as little $3.50 per ous work. One possible cxplanalion is that legal suppliers gram in transactions of 100 g. The price per gram for black incur tax and regulaioiy cosis that are evaded or avoided by market heroin transactions that are roughly 100 pure grams black markel suppliers. In this seciion. I examine ihe mag- is $113. Assuming that heroin is three times as potent as nitude of such costs. morphine (Krivanek, 1988. p. 106), this implies that black The most obvious cost evaded b> black market suppliers market heroin is 13 times the price of legal heroin.'' is taxes and associated compliance costs. As.suming no barriers to labor and capital mobility, the differential in C. Caveats labor and capital costs between the legal and black market sectors equals the tax burden incurred by lab(.>r and capital One possible qualification to these results is that ihe black in the legal .sector. market prices are for goods whose quality might be low or In 1995, tax collections in Ihe relevant categories equaled uncertain. It is not possible to eliminate this concern, but 26% of GDP (Miron, 2001, pp. 25-26). Adjusting for tbe several considerations suggest it is not of overwhelming difference between gross and net output, or for the fact that importanee. First, the analysis above controls for purity, the capital supplied to the and nonprofit sectors is single most important aspect of quality. Second, black not taxed, would make the lax burden even larger. Likewise, market suppliers have an incentive to provide quality to allowing for compliance costs would increase the estimated attract repeat business. Third, existing data suggest quality burden on legal suppliers. considerations are not overwhelming. A second source of cost-increasing policies is environ- mental, safety, and health regulation. These policies impose ''Thiscalctiiation ignores any to.sts of producing heroin iVoni niorphiiR-. Maher (1976. p. 40) states thai "heroin is approximately 4-M times more '"The reason given for about id'i i^'i ihc episodes is "other/iinknow]! potent than morjihine when admini.stercd hy injcclion;' which woiild reason. Distribiuing ihese cases proponionally amongst ihe renuiining imply a lower raiio of black markel to legal heroin |iiiees. five calegories would not have a snbstanlial etfecf on the ivsiills. 528 THE REVIEW OF ECONOMICS AND STATISTICS direct costs in the form of compliance expenditures, and niunication) suggests Ihat $10 an hour is an upper bound on they potentially reduce productivity growth by altering tirni the wages paid to employees in the illegal drug sector; input choices or reducing innovation. Direct compliance higher figures sometimes cited in the literattire (lor example, costs are modest in comparison with total firm costs (Ber- Fagan, 1996) correspond to the earnings of entrepreneurs, man and Bui, 1999), but existing estimates suggest this who put up their own money,^- Since many black market regulation has reduced productivity growth by at least employees simultaneously work at minmium-wage Jobs 0.10-0.20 percentage points per year and possibly by as (Lcviit and Venkatesh, 1998). the niininunn wage is a lower much as 1-2 percentage points per year (Miron, 2001, pp. bound on the earnings of (hese employees in the legal sector. 26-28). Thus, wage ct)sts are approximately double what they would Assuming reduced productivity growth of 0.5 percenlage be hi a legal market.'* points per year cumulated over twenty-five years {the period Prohibition also imposes costs on the owners of capital. from the early 1970s, when this regulation began to bite, to Both physical and fmancial assets believed to have been the present), this regulation has reduced productivity, and used in the drug trade are seized by federal, .suite, and local thus increased prices, by 13.3%. Assuming the manufactur- law enforcement authorities, and a substantial fraction of ing price is 75% of the retail priee implies thai environmen- these assets are forfeited permanently. Over the period tal, safety, and health regulation have increased retail prices 1993-1997, seizures averaged at most $2.5 billion per year by about 10%." (Miron, 2001, pp. 32-33). A different cost avoided by black market suppliers is To gauge the impact of these seizures on costs. 1 assume advertising. The implication of these expenditures for cost as a benchmark that the capital-output ratio in the drug and price is ambiguous a priori (Miron, 2001). But any sector is equal to the ratio of nonresidential private eapital to reduction in conventional advertising offsets increased costs private output in the overall .'"* In 1996, this was necessitated by this reduction (for example, violent airf approximately 1.25 (U.S. Census Bureau, 1998, Table 890, battles), and it may imply price reductions that exceed the p, 559: Table 716, p. 452). ONDCP (1997a. p, 3) estimates magnitude of the reduced expenditure. The scope for adver- that 1995 sales in the illegal drug sector equaled S57.3 tising in a lega! market for cocaine or heroin might resemble billion 1996 dollars. Assuming a 1996 value of $60 billion, that in the markets for soft drinks, alcohol, or cigarclles; this implies a capital stock of $75 billion. Thus, asset advertising accounts for about 10% of revenues in the soft seizures in 1996 equaled approximately 3-3% of the capital drink industry, 7%-15% in the alcohol industry, and 12% in stock. the cigarette industry (Miron, 2001, p. 29). The third major cost imposed on drug traffickers is drug Numerous other policies (minimum wage hiws, antidis- seizures, ONDCP (1997a, p. 4) estimates that in 1995, crimination laws, collective bargaining laws, fees and per- 462-553 metric tons of CHCL were produced and destined mits, local zoning rules, liability insurance, and the like) for the Linited Slates; of this amount, foreign seizures almost certainly add a few percentage points more to the amounted to 41 tons and federal seizures 98 tons, implying costs evaded or avoided by black market suppliers. Adding a seizure rate of 25%-3O*/r. Adding in stale and local up these factors implies that legal price exceeds costs by a seizures would likely increase the seizure rate only moder- factor of roughly 2. ately. To determine the net impact of these three effects on the costs of producing drugs under prohibition, assume lirst that V. The Costs Imposed by Prohibition Enforcement seized diues are raw materials. Then write costs as

Although black market suppliers evade many costs in- C = \vL + rK + c/M. (I) curred by legal suppliers, they incur costs not levied on legal suppliers. In fiscal year 2000. federal expenditure for pro- where C is the total cost. \v is the wage rate, L is labor, r is hibition enforcement was approximately $11.7 billion, the rental rate on capital, K is capital, (/ is the price of raw while and loeal expenditure almost certainly exceeded materials, and M is the quantity of raw materials. The $14.3 billion and may have been as high as $50.2 billion discussion above suggests that enforcement raises vr by a (Miron, 2002). This might suggest that prohibition has a factor of about 2: that it raises the required rental rate on substantial effect on drug prices. I show here, however, that capital by }.^ percentage points, which is roughly 50% of the costs imposed by enforcement are consistent with rela- the average real return on equity; and that it raises the tively modest effects of prohibition on prices. The first cost imposed by prohibition is the wage pre- I- t.cvill and Vcnkatesli I I9<-)81. Padilla i 19921. and Rourgois I 199?) also mium paid to compensate employees for the risk of arrest, rcpoii linv wages fur workfis in tiie bla^k niailci drii^ lr;idc. " This Loncliisioii is consisicni with indeiiondciU evidL'ncc in Macfoun incarceration, injury, or dealh. Levitt (1999, personal com- and keuicr MW2). Lott (1992). Gniggi^-r (IW'^). and Khng (IW9|; SLV Mimn (201)1. pp. 31-.^2). II Nevo (1998) estimates thai the matiufaciuring prict is ol ihe '•' I his is a stronii assuinplion. hui ii is nut clear wiiicli way il is I.-Mascd\ retail price for ready-to-eai cereals. Bulow and Kicmpercr in sonic cases black market suppliers rciy hea\il\ on capital to evade huv 58% for cigarettes. enrorcenienl. DRUG PRICBS 529 effective price of raw materials by no more than 50%. same time period. This combinatitin of facts is a fertile topic Combining tiiese estimates implies enforcement raises the for future research. price by at most a factor of two. Alternatively, assume that seized drugs are the finished product. Then costs are REPIZRENCliS Basov. Sureii. Mircillc .lacohson. and .lelfrcy A. Miron. "Prohibition and C= lwL+ rK)(\ + t). (2) the Market lor lilegal Drugs: An Overview of Recent History." World Economics 2:4 12001'). I.WIS8. Bernian. Eli. and Linda'I'.M. Bui. ""Hn\ironmenUil Regulation and Pro- where / is the seizure rate. In this case, the estimates imply ducliviiy; Hvidencc Irom Oil Refineries," Boston University manu- that enforcement raises costs by at most a factor of three. script (1999). These costs are roughly the magnitude of the tax and Bourgois. Philippe, hi Srarrh oj RcspvcL Selling Crack in El Barrio regulatory costs avoided by black market suppliers. (Cambridge. MA: Cambridge University Pres.s. 1995). Bulow. .lereiny. and Paul Kleniperer, "The Tobacco Deal," Brookings The full effect of prohibition-cum-enforcement is pre- Papers on Fcoiiomiv Acliviiy. Microeconoinits 1998. 323-394 sumably larger than the factor of 2-3 suggested above, since (IW9). Callerall, William A., and Kennelh Mackie, "l.ocal Anesthetics." in Joel the calculations ignore additional costs necessitated by pro- G. IlardiTian and Lee H. Limbird (Eds.), (loodmiin .i.sofl herapi'uriis, 9f/!('(y.( New York: McGraw- factor proportions or economies of scale. It is not obvious Hill. 19%). how large these additional effects might be, so the estimates Ciiulkins. Jonathan P.. and Peier Reuier. "What Price Data Tell Us about Drug Markets," Joimial oJ Dn/g AvM/t'.v 28:3. 593-612 (1998). in this section merely suggest that the tangible effects of CoffeeUniverse. ltUp://www.eolTeeuniverse.com/university.html (1999). enforcement are not so large as to render implausible the i^iNardo. John, "Law Enforcement. Ihe Price of Cocaine, and Cocaine results above about legal versus illicit prices. Use," Mathcmciticii! and Conipulcr Modelling. 17:2. 53-64 11993). l-agan. Jeffrey. "Gangs. Drugs, and Neighborhood Change," in C. Ronald VI. Conclusions Huff |Hd,). Cniiigs ill America (Newbury Park, CA: Sage Publica- lions. 1996). The analysis above suggests that the cunent prices of Federal Register. htip://t nAebgate3.access.gpo.gov/ctii-bin/waisgate.cgi? WAISdocli:)-38967911899 + 5+() + ()& WAI.Saction=relneve(l998). cocaine and heroin, though substantially higher than they i-riedman. Milloii. "Prohibilion and Drugs," Newsweek (May 1, 1972), would be in a legal market, are not as elevated as suggested 104. by previous research. Taking into account both the fanmgale- Grinspoon. Lesler. and James B. Bakalar, Cocaine: A Drug and Its Social iWolmion (New York: Basic Rook.s. 1976). to-retail price differences and the legal-versus-illegal price Grogger. Jeffrey, "The liflect of Arrests on die Employment and Barnings comparisons. I estimate that the black market price of of Young Men," Qmirleiiv Journal of Economics CX:1 (1995). cocaine is 2-4 times the price that would obtain in a legal 51 72. International Cocoa Organi/alion. Web site, http://www.icco.org/ (1999). market, and of heroin 6-19 times. In contrast, prior research Kling. Jeffrey R.. "Tbe Effecl of Prison Senience Lenglh on tbe Subse- has suggested that cocaine sells at 10 to 40 times its legal L|uenl Employment and Earnings of Criminal Defendants," Wood- price and heroin at hundreds of times its legal price. Thus, row Wilson Scbooi discussion paper no. 208, Princeton University (1999). my conclusion is qualitatively similar but quantitatively Koper. Chrisiophcr S.. and Peter Reuier. "Suppressing Illegal Gun Mar- different from thai in previous research. kets: Lessons from Drug Hnforcemoiil," hiw and Contemporary The key question for future research is whether increased Prohlew.'. 59.] (1996), 119-146. Krivanek. Jara. Heroin: Mvlhs and Rcuinv. Sydney: Allen & Unwin enforcement of a given prohibition raises prices, whatever (1988). the relation between the initial level of prices and those that Levin. Steven iX, and Stidhir Alladi Venkaicsb. "An Eeonomic Analysis would obtain under an altemative regime. The discussion of a Dnig-Selling (}ang"s Finances." NBER working paper no. 6.592 (1998). above does not address this issue; moreover, it seems likely Lewis. Roger. "The Illicil Traflic in Heroin: Introduction and Part 1, that increased enforcement should add costs and therefore Cultivation and Produclioii." Dniglink (Spring Iy84), 7-14. increase prices, independent of the considerations raised Lott. John K. Jr.. "An .^ltempl ai Measuring the Total Monetary Penalty from Drug Convielions: The Importance of an Individual's Repu- here. tation." ,^»n;

for Allcnuilivc.s (Stanford: Hoover Institution Press. Crime imd .hislici': Ait .Anniiid ficivcir of Researeh. Vol. 7 (Chi- IWl cago: University ot Chicago Press. I9K6). Nadelmaiin. l:th;m A.. "America's Drug Problem." Biilh-rin of tlw Aim-r- Stecklow. Ste\e, and Jonathan Karp. "Poppy Kields of India Have a Placo ivan Academy oJ'Ans and Sricncc.s. XLV:.i 1199!). 24-40. in the Sun. linl It's under .Atlack," Wall Sireci Journal (.April 3. Ne\o. Aviv. "Measiirins Maikct Power in tlit; keady-lo-Hut Ceieal Incitis- try." NBEk vvor'king paper no. 63