Dealing with Drug Abuse

A Report to the Ford Foundation

THE DRUG ABUSE SURVEY PROJECT

STAFF PAPER 4

The Economics of Heroin

by John F. Holahan, with the assistance of Paul A. Henningsen

Production * Policy Options: What Can Be Done * Distribution * Consumption

This paper applies economic analysis to the heroin market and to the effect of alternative government policies on this market. The principal concern is supply, but some discussion of demand is included because the effects of public policy on the supply of drugs cannot be adequately analyzed without some knowledge of the effect on demand of changes in price, risk, and availability. The analysis must be considered tentative. Accurate information on the economic aspects of all drugs of abuse is scarce, that which exists is frequently unreliable, and conclusions are often suspect. This paper is no exception.

Much of the data used were taken from two studies: (1) The World Opium Situation (October, 1970), by the Bureau of Narcotics and Dangerous Drugs (BNDD) Intelligence Staff; and (2) Economics of Heroin Distribution (1970), by Mark Moore. While these papers contain the most extensive economic research available in this area, the information is not unimpeachable. This is not because of deficiencies in the studies-which are excellent -but because of the inherent difficulty of gathering accurate information on illegal enterprises or consumers of illegal commodities.

The BNDD staff paper is the result of an extensive survey of opium production and consumption throughout the world. The paper provides considerably more data, in terms of both quantity and quality, than are available from the United Nations or from other publications. Unfortunately, it is not possible to verify this information through other sources, nor does it appear that it will be possible to do so in the near future.

The Moore paper provides a macroeconomic analysis of the heroin distribution system, as well as considerable interview data on daily expenditures of addicts and sources of funds. The data are not derived from sufficiently large samples to allow precise estimates, but they do indicate rough orders of magnitude. Other data derived from interviews with BNDD staff members or U.N. publications are subject to the same general weaknesses.

PRODUCTION

SOURCES OF SUPPLY

The production and distribution of opiate narcotics at the international level are complex and little-understood matters. The crisis of heroin addiction that has afflicted the United States in recent years has obscured the fact that the growing of opium is an age-old, legal occupation in many areas of the world. Heroin addicts in the United States consume only a very small amount of the total quantity of opiates produced worldwide. The Bureau of Narcotics and Dangerous Drugs estimates that 10,000 to 12,000 pounds of heroin are sufficient to supply the entire addict population in the United States for one year.' Since, as is generally acknowledged, ten pounds of opium are required to manufacture one pound of heroin, only 100,000 to 120,000 pounds of opium are needed for illicit heroin use in the United States about 2 per cent of the estimated world production .2

Of the opium used by the U.S. heroin market, 80 per cent comes from Turkey, 15 per cent from Mexico, and 5 per cent from the Far East.' These are not precise figures, however, but assumptions rather generally derived from heroin seizures and knowledge of the international illicit drug-import routes.' Moreover, several recent developments may change the sources of supply for the U.S. market. On June 30. 1971, President Cevdet Sunay of Turkey decreed that all poppy cultivation and opium production in that country will be forbidden beginning in the fall of 1972. Until that time, licit opium production will be limited to only four provinces and strictly policed. How well this policy will be enforced, and what ultimate effect it will have on heroin use in the United States, cannot as yet be determined. However, already there has been considerable speculation that the Far East supplies a greater amount than was thought previously. The tighter controls now being imposed by Turkey and the increased contacts between the U.S. and Southeast Asia caused by the Indo-China War have shifted the sources of supply somewhat.

Raw opium itself is produced from the poppy plant (papaver somniferum-the sleep-bearing poppy), which consists of a thin main stalk, three to four feet tall, with several egg-shaped pods on top. Between June and July, the poppy blooms, and, 10 days later, when the blossoms lose their petals, the pods are ready for the opium resin to be extracted. Incisions are made in the pods, and a milky fluid oozes out. At the end of the day, this substance is scraped off the pods by hand. It later hardens into a brown gum and is pressed into cakes of raw opium. The amount of opium produced from a poppy crop will increase if this process is repeated once again, as in Turkey, or several times, as in India, but the morphine content will be reduced with each lancing.'

Many factors affect the opium yield per acre-the quality of the soil, the temperature, rainfall, the quality of the seedirriga-tion, fertilization, and crop rotation. Intensive cultivation and a great deal of labor are required to produce opium, particularly during harvest, when the pods must be incised and the raw opium collected by hand, but labor is also needed during the growing season to thin the poppy plants. It is estimated that between 175 and 250 man-hours of labor are required to produce 1 kilo of opium.6 Because of this labor-intensive production process, poppies tend to be raised only where labor is abundant and relatively inexpensive; but there seems no clear reason, given a different set of relative prices for labor and capital, that a more capital-intensive process could not be employed.

Most poppy cultivation occurs in a zone stretching from central Turkey through Iran, Afghanistan, Pakistan, India, Burma, Laos, Thailand, and southern China, with the greatest concentration in India and Southeast Asia. Poppy cultivation is legal in India, Turkey, the U.S.S.R., and Iran. Turkish licit poppy acreage was reported to be 29,600 acres in 1970.1 According to U.N. figures, India has at least 86,500 acres planted.' Iran, which abolished the production of opium from 1956 to 1968, planned 29,600 acres of poppy cultivation for 1970. In addition, roughly 30,000-32,000 acres each are believed to be planted in the Soviet Union, Afghanistan, and Pakistan, and negligible amounts of licit opium are produced in Yugoslavia and Japan.'

According to figures reported to the United Nations by producing countries, the total licit world production of opium is 1,0601,085 metric tons, or approximately 40-45 per cent of the total produced. India is by far the world's largest licit producer of opium, with a reported 750 metric tons in 1968. Turkey and the U.S.S.R. each contribute 115-120 metric tons of licit production. An estimate of 75-100 metric tons was made for Communist China, based on expected medicinal needs for its population. The estimated totals for licit and illicit production are presented in Table 4-1.

Licit opium is used primarily in the manufacture of medicinal opiates. Legal morphine production currently amounts to approximately 150 metric tons per year-a substantial rise from around 85 metric tons per year in the mid-1950's. This growth is due chiefly to the rising demand for codeine, but a small portion of the licit opium is used for opium-maintenance programs in India, Pakistan, and Iran.

The world's illicit production of opium is estimated at 1,2501,400 metric tons annually.10 The principal concentration of illicit production is in the Burma-Laos-Thailand region, which
produces an estimated 700-750 tons. The Afghanistan-Pakistan area produces another 250-300 tons of illicit opium, and the illicit output of Turkey is estimated at 100 tons. Illicit cultivation of the plant also occurs in parts of Mexico, South America, and North Africa, although the amounts produced there are negligible. While some illegal acreage is planted in Turkey, most of the illicit opium there is diverted from the yield on legal acreage. On the other hand, all the opium produced in Afghanistan, Burma, Laos, and Thailand and part of the outputs of India and Pakistan are derived from illegal acreage. About 40 metric tons of illicit Turkish production was thought to enter Iran in 1970; the remainder is exported to Western Europe and the United States.

The Southeast Asian countries consume most of their own opium, but some is shipped to Hong Kong for use there and for distribution throughout the Far East. A small amount, perhaps one or two metric tons, is shipped to the United States, although this figure has probably risen recently. India also consumes the major part of its illicit production; and, although some of the Afghanistan-Pakistan production is consumed internally, approximately 200 metric tons is thought to enter Iran. Iran is the largest consumer of illicit opium in the world, with an estimated 350,000 opium and heroin users. In 1969, Iran began controlled production of opium in order to maintain its own addicts.

YIELDS, MARKET PRICES, AND INCOMES

Income from poppy cultivation is affected by the opium yield obtained, the morphine content of the raw opium, and the current licit and illicit market prices. The reported yields per acre in 1969 for four countries are presented in Table 4-2, but, since these are government figures provided by the local farmers, both yields and acreage are probably understated. This is undoubtedly the case for the crop yields in Turkey and Pakistan. Turkish yields are closer to 13.4-14.2 pounds per acre, with the nonreported amount diverted to illicit traders." Opium yields are estimated at 8.6-28.6 pounds per acre in Southeast Asia" (probably closer to the lower figure"), and the Soviet Union reports yields of 27.5-32 pounds per acre."

Prices for opium vary with quality. The morphine content of Turkish opium, reportedly the highest in the world, varies from 9 to 13 per cent, while in other producing countries the morphine content varies from 4 to 12 per cent. Prices to farmers for raw opium are presented in Table 4-3. Because of its quality, the prices for Turkish opium are the highest, with the exception of the Iranian price, which is supported by the government in order to discourage diversion to illicit traders. The licit prices are determined by the world market price: the price at which the national opium monopolies can sell to pharmaceutical companies. In Turkey, the illicit price varies with the yield of the current year's harvest, the season, enforcement efforts, and the demand for Turkish opium from other countries, particularly Iran. The illicit price in Iran is extremely high because of the strict enforcement of laws against illicit trading.

The absolute income per farm from opium cultivation is greatest in Turkey, but the percentage of total farm income accounted for by opium increases as one moves eastward. In the late 1960's, there were an estimated 70,000 farms producing opium in Turkey. If 120 tons are sold by these farms at the licit price of $11 per kilo 0 kilo=2.2046 pounds) and 100 tons on the illicit market at $25 per kilo, the gross income from opium amounts to $3,820,000, or $55 per farm. This amounts to roughly 5-10 per cent of farm income in Turkey. In India, roughly 200,000 farms earn $70-75 from poppy cultivation, or 15-20 per cent of average total income per farm. Since the income figures include income-in-kind (food, clothing produced on farm), the income from opium provides an even higher percentage of cash income than we have indicated. Opium is probably the principal source of farm income in Burma, Laos, and Thailand; cultivation is said to be concentrated in areas controlled by guerrilla insurgents, who transport much of the output to Hong Kong and use the proceeds for arms purchases.','

Licit production of opium is of minor importance even in India and Turkey. India earned roughly $6-7 million for exports of 530 metric tons in 1968, compared with total export earnings of $2 billion. Indian farmers earned approximately $7.7 million on production of 700 metric tons. Turkish exports of 120 metric tons yield approximately $1.7 million, less than 0.3 per cent of total exports. Turkish production is also very small in relation to national income of $9 billion. On the other hand, opium sales are undoubtedly of great economic importance to the producing regions of Burma, Laos, and Thailand.

POLICY OPTIONS: WHAT CAN BE DONE

Policies designed to attack the heroin problem at the source of supply devolve into two basic alternatives: (1) the purchase of entire opium crops, or (2) direct controls on production. Each of these proposals has problems difficult to surmount at present.

PRE-EMPTIVE BUYING

The first alternative would be for the United States to purchase the total output of opium from those countries that supply the heroin entering its borders. The administrative simplicity of this proposal makes it appear deceptively attractive. In time, however, pre-emptive buying would necessarily involve purchases of both licit and illicit production in all countries where opium is currently grown. The U.S. offer price would have to be raised above the maximum price illicit traders are willing to pay; otherwise, the farmer would have an incentive to sell to the higher bidder. If the offer price is less than the illicit price, leakages into the illicit market would probably result, and the policy would therefore have little effect on the supply of opiates sent to the United States, especially since the amount is so small relative to total world production. And, seemingly, the illicit trader would be willing to increase his offer price substantially, because, given the large markups at each stage of the distribution system, the price of raw opium has relatively little effect on the ultimate retail price in the United States (see Table 4-4). (For example, if the current price to the farmer for 10 kilos of illicit opium is $250, and the retail street price for I kilo of heroin is $264,900, then raising the farm price 100 per cent would raise the retail price only .09 per cent.)

A pre-emptive buying policy would become very costly, because opium producers would have a clear incentive to increase their outputs, and other farmers would have a similar incentive to begin substituting opium for other crops. Eugene Rossides, Assistant Secretary of the Treasury for Enforcement and Operations, has gone on record as strongly opposing any concept of pre-emptive buying. He told a congressional committee that "it would simply stimulate production and it would take away the responsibility of each nation to handle the problem as part of the international community.""

Moreover, without strict administrative controls, this policy would not only increase production in all countries currently producing opium but might induce other countries whose climate is suitable to begin production in the hope of sharing in the windfall.

If this policy were directed at only one source of supply, such as Turkey, it would not prevent traders from obtaining illicit supplies in other producing countries. The 100,000-120,000 pounds of opium used by the U.S. market could be grown in an area of, at most, 11.2-23.2 square miles and thus could easily be absorbed in another producing region." In fact, some experts believe that 35 pounds of opium could be produced on a single acre of land if a capital-intensive process were employed under controlled circumstances. -If this proved true, -generally, only 4.55.4 square miles would be needed to supply the entire U.S. market. Consequently, a pre-emptive buying policy, in the long run, could result in a change of location for the source of U.S. supply with little or no effect on the addiction problem.

Pre-emptive buying within an individual country might prove effective in dealing with an internal addiction problem if coupled with strict limitations on acreage and illicit importation. Iran, for example, has discouraged export to illicit markets in other countries by purchasing its entire internal opium crop at a very high price, restricting production to certain designated areas, and strictly enforcing its laws against illicit markets. The legal opium purchased by the government is used by the state-run addict maintenance program, in which 50,000 individuals are registered. The potential expense is one problem with the high-support-price approach. If nearly all Iran's 300,000-400,000 users and addicts entered the program, the cost would rise dramatically. Furthermore, even with severe penalties against smuggling and illicit trading, Iran still experiences difficulty in trying to prevent the importation of opiates. Some success has been reported in eliminating illicit supplies from Turkey, but imports from Afghanistan and Pakistan have not been halted. The mid-1970 price in Iran for Afghani opium was well below the licit price, despite intensive law-enforcement efforts. It seems that the heavily armed mountain tribes operating out of both Afghanistan and Pakistan have not as yet been deterred by either the border authorities or the death penalty on smuggling. On the other hand, the price of illicit Turkish opium in 1970 was as high as $400 per kilo in Tehran-six to seven times the government support price.

As the Iranian example shows, the success of a pre-emptive buying policy, even if accompanied by administrative controls, depends on whether government can effectively exert its will over the producing regions. In fact, most of the world's opium supply is produced in areas where little political control exists. Administrative controls themselves are difficult to enforce, as we shall see. Therefore, because of the size of the world market relative to U.S. heroin consumption, the high elasticity of supply, and the difficulty of control, it is unlikely that a policy of pre-emptive buying could be successful.

DIRECT CONTROLS

The application of administrative controls alone is the other basic alternative to reduce opium supplies. This involves restricting the acreage on which opium can be grown and maintaining a low official price. A low official price may reduce the illicit price, at least in the short run, and thereby discourage illegal planting and crop diversion. Certainly, taken by itself, a low official price will not induce new production. But the most salient feature of direct controls is increasing the risks of illicit trading and consequently reducing the profitability.

Practically, however, continuous enforcement of direct controls is both difficult and costly, and the necessary resources available to most producing countries are quite limited. Especially in countries with weak central governments, controls present severe administrative problems. All countries with an illicit trade lack adequate numbers of both trained law enforcement personnel and motor vehicles. Given the low wages in public-service jobs and the large profits in the opium trade, corruption of government and police officials is a widespread problem in most countries and is likely to continue. The bribes that illegal importers are willing and able to pay are considerably higher than the amounts officials are currently accepting for inaction or nonenforcement of the law. Widespread social acceptance of opium use in many rural areas of Pakistan, Afghanistan, India, Burma, Laos, and Thailand is another difficulty that makes enforcement of controls for the benefit of an outside country appear especially burdensome. In Turkey, where there is reportedly little opium use, controls have been extremely unpopular because of both the loss of an important source of farm income and the rising anti-American sentiment. While limitations on local financial resources could presumably be eased by American aid, the administrative and political difficulties severely limit the possibility of successful controls.

If, however, the desire and resources to do so were present, controls could probably be applied successfully in certain individual countries. Communist China and Iran successfully eliminated production in the 1950's. Turkey has restricted legal acreage of poppy cultivation to 29,652 acres from 103,782 acres in 1960. In the process, illegal production may have been somewhat reduced, because illicit Turkish opium is believed to come principally from larger than declared yields on legal acreage." Turkish officials seem committed to their pledge to enforce a ban on all poppy cultivation by the fall of 1972. The chances of this policy's success will be discussed later.

The major problem in curtailing opium production is that most of the world's supply is produced in areas where no national control exists. The lack of control is virtually complete in the producing regions of Burma, Laos, and Thailand. Areas that are largely controlled by tribal groups are responsible for opium production and distribution in Pakistan and Afghanistan. Therefore, it seems likely that attempts to control or abolish opium production in Turkey would have little effect, as illicit traders turned to these areas for supplies. Indeed, the recent changes in Turkey do not seem to have caused any decrease in total U.S. imports of illegal opiates.

Still another difficulty is that of restricting opium production in one country when neighboring countries do not, or cannot, restrict production. After twelve years of prohibition, Iran returned to limited production, because its addicts were purchasing illicit opium from Turkey, Pakistan, and Afghanistan.

Even in Turkey and India, control systems have failed to halt the cultivation of illicit opium. Turkish farmers make written declarations to the government of their licensed acreage and expected kilogram yields, selling the surplus to the illicit market. In 1969, for example, Turkish officials collected 116 tons of licit opium from poppies in eleven provinces. But in 1970, when nine provinces were allowed to produce, only 60 tons were turned in. Presumably diversion accounts for some of the difference. Some illicit production in Turkey has occurred from unlicensed acreage, but this is relatively easy to discover if sufficient efforts are made, since the poppies have a long growing season and are easy to detect, especially from the air. In India, on the other hand, the illicit opium output is produced principally on unlicensed acreage.

Direct controls might be more effective if they were combined with crop substitution, subsidies for production of other goods, or direct payments for not growing poppies. But crop substitution is unlikely to be successful, since it is difficult, if not impossible, to find crops that yield as high a return per unit of land as opium. For example, attempts to substitute a high-yielding Mexican wheat in Pakistan failed because the crop provided a return per acre roughly half that of opium, assuming opium is sold at official prices."

In Turkey, there has been a sound agricultural reason that crop substitution is difficult. Apart from low-income-producing grains, the poppy is the only autumn-sown crop and is not easily replaceable by the higher-yielding spring crops, such as cotton, maize, sugar beets, tobacco, or melons. The opium produced is important to the small farmer, not only because of the amount of income gained, but also because it is harvested before the grain and therefore provides expense money for the main harvest. Furthermore, the harvest of opium occurs in a period when the farmer has little else to do and can spend his time on the lancing of the poppies and collection of the opium.

Subsidies for production of alternative crops would certainly be required if the substitute crops provided less income than opium. The cost of producing a unit of a substitute crop would be reduced by a subsidy, and the transfer payment from the government would allow the grower to increase production and, in the process, bid resources, land, and labor away from other crops, including opium. Because these inputs would become more expensive, the cost per unit of producing opium would increase.

However, if demand for opium in this particular national black market is relatively insensitive to changes in price, opium production would be little reduced, and little reduction in farmer income would result. A higher subsidy would be necessary to raise the relative return on the alternative crop. If the demand for opium varies with price," perhaps because of the availability of opium from other sources or synthetic opiates, then the subsidy would probably reduce much of the country's production.

Clearly, subsidies for output of other goods could be set high enough so that the output of opium in a given country could be greatly reduced. This may come, however, only at great expense to the government or to the United States, if the latter sponsors the program. The same types of administrative problems that exist in the crop-purchase proposal are repeated here.

Basically, it may be difficult to determine how much of the production of the subsidized crop is actually land diverted from opium.

Similar administrative problems would result if governments made payments to producers for not growing opium, since it would be difficult to determine how many farmers who do not grow opium now would demand payment. In 1972, when Turkey's complete prohibition on poppy cultivation goes into effect, the government plans to pay farmers on the basis of the value on the international market of their whole opium crop sold the previous year to state officials. But, if only those who were originally "licensed" to produce are given payments, and no payments go to those who planted illegal acreage, there is no incentive for these growers to cease producing opium, nor is there an incentive for nongrowers to resist the urge to produce. Furthermore, to be effective, payments must exceed the return on output sold illegally. Since the illicit price of opium is extremely low relative to the retail price of heroin in the United States, it could rise greatly without much effect on the final price to the addict. Therefore, payments would have to be extremely high to be effective and would make the program very costly. Thus, it also seems unlikely that a policy of direct controls on opium production, even if combined with crop substitution, subsidies for production of other goods, or direct payments for not growing opium, could be completely successful.

General economic development in poppy-growing areas may eventually reduce the relative profitability of this crop. The usual process is for technological innovations in agriculture to increase the productivity of fanns to the point where large surpluses of labor develop. Migration of labor to the cities occurs because of the availability of employment and public services. As industrial development increases the demand for labor, real wages rise, and capital-intensive production processes become more desirable. If no capital-intensive method of cultivating opium can be found, then the return on poppy falls relative to other goods and opium cultivation is greatly reduced. Clearly, this ignores changes in the demand for opium; moreover, it is, at best, a long-run solution. This situation may eventually develop in Turkey but is not likely for a very long time in the East.

DISTRIBUTION

Several books and articles have described the system ky which opium is transformed into heroin and imported into the United States. They are all in substantial agreement, possibly because the primary source of information on the subject is the Bureau of Narcotics and Dangerous Drugs.

In the case of Turkish opium, growers either withhold part of their output or plant illegal acreage. Small-scale entrepreneurs gather opium from the farms in relatively small quantities and deliver it to syndicates in or near Istanbul. At this point, the opium is transformed into morphine, which cuts the volume by 90 per cent. This process formerly took place in Syria. but as relations between Syria and Turkey have deteriorated in recent years, border controls have been tightened and the morphine labs moved to the Istanbul area. From Istanbul, the morphine is transported overland through Bulgaria and Yugoslavia or by ship to its destination in southern France, where covert laboratories refine the morphine into heroin. (One theory is that some of the heroin labs are located in trucks and are thus mobile and difficult to detect.) Once refined, the heroin is smuggled to the United States through New York, Montreal, or some other large port. According to some sources, shipments are also being sent to the U.S. via South America through a Southern port.

MARKET STRUCTURE

The best available outline of the domestic heroin distribution system, at least for New York City, was developed by Mark Moore of the Hudson Institute." His outline relies, in turn, on earlier analyses by John Casey and Edward Preble" and information from the Bureau of Narcotics and Dangerous Drugs.

Moore suggests that heroin passes through six levels of distribution: the importer, the kilo-connection, the connection, the weight dealer, the street dealer, and, finally, the juggler or pusher. Table 4-4 shows the outline of the system and includes estimates of the prices and value added to the heroin at each level. Before reviewing the operation of the specific segments of the system, it is important to discuss the over-all market structure of the business. The key questions here are: How tight is vertical organization? How competitive is the system at the various levels? How much increase in risk would be required to make it unprofitable for large operators to stay in business?

For many years, conventional wisdom held that the heroin business was monopolized by organized crime, perhaps by the Mafia. BNDD now seems to believe, probably as a result of recent arrests of South Americans at relatively high levels, that 10 or 12 large wholesalers with varying degrees of connection to organized crime now control the distribution process in the United States.

It seems clear that many economic principles operating in society at large must also apply to illegal enterprises," since the business of supply and distribution of illicit drugs is presumably subject to the same general economic laws, and responds to the same general set of incentives, as prevail in more open industries. There are, of course, some crucial differences caused by the fact of illegality and also by the moral problems associated with dealing in many of these drugs-for example, the need to minimize the risk of detection of the origin of the distribution system.

The Hudson monograph argues that the objective of the industry is to maximize profits within the constraint of the risk of arrest and imprisonment, and that this objective is most likely to be attained if the industry is organized into several small distribution units separated vertically from one another, if higher levels in the system are monopolized, and if lower levels are competitive with some elements of market power (monopolistic competition). The small distribution units in the heroin industry are separated from one another by restricting information at one level about the identity and activities of other levels. Risks are limited by close supervision and discipline of employees, thus reducing the probability of information leaks, and by the fact that penetration of one level will not result in penetration of another. The primary disadvantage of small, separate distribution units is their inability to adjust to large changes in demand; suppliers end up with either excessive or deficient inventories.

Moore argues that a monopoly could most efficiently perform the functions required of the high levels in the distribution network. The top levels in the system must: (1) restrict the total supply of heroin to maintain high market prices, (2) regularly and reliably supply an amount of heroin fairly close to the realized demand, (3) manipulate supply conditions above them with a minimum of explicit planning and negotiation, and (4) adjust to errors in supply with a minimum of negotiation and activity. A centralized organization would be more effective in handling each of these problems. A monopolist can restrict the total supply, allowing high prices throughout the system. In addition, he is in a position to have an efficient ordering policy, thus successfully accommodating uneven deliveries and lead time. Agents can be coerced into holding inventories or resupplied when short of heroin. Tight control can be maintained over the circulation of information. Finally, a monopolist can maintain a monopoly position in relation to producers.. thus reducing lead times, unevenness, and quality deterioration.

In contrast, a competitive market would be less likely to ensure a smooth, restricted supply, adjusting to errors in a discreet, efficient manner. In a more competitive situation, it is more difficult to pool shipments or parcel out uneven deliveries of heroin. With many suppliers at the top, there are more transactions, more explicit negotiations, more occasions of short supply or large inventories, more information circulating freely-and thus a greater probability of arrest. A monopoly, therefore, is logically the most efficient type of organizational structure at high levels.

But a successful monopoly requires the existence of some type of leverage or barrier to entry, such as (1) special knowledge unavailable to others, (2) total control over some vital raw material, or (3) total control over the market as to make the cost of entry prohibitive, especially if entry requires a large fixed investment. It is doubtful that any of these considerations applies to the heroin traffic. The knowledge required to refine opium into heroin is certainly not difficult to acquire. It is also doubtful that the knowledge of methods of conducting illegal activities-such as smuggling-is so limited as to allow any firm to maintain a monopoly. There are no substantial fixed costs other than the investment in an information system, needed to offset the high level of risk in the trade. However, the difficulty, and thus the cost, of developing such an information network may be a formidable barrier to entry.

Trusted personal relationships with Turkish or French syndicates may have put some organizations in a monopoly position for Turkish opium. However, it would be in the interest of the Turkish or French firms to sell to as many importers as possible, forcing them to bid against each other and thus raise the price of heroin. A firm desiring to enter the business should, therefore, be able to establish the necessary connections.

It is possible, of course, that the suppliers are not profit maximizers; they may not wish to accept increased risks for the sake of increased profits. Instead, after establishing an acceptably low level of risk, they may seek to maximize profits within this constraint. Such an approach would sacrifice many of the benefits of competition among buyers because of the need to limit the number of people who know the supplier's identity. (This factor also operates as a barrier to entry at other high levels of the distribution system.) If the barriers to entry in the market for Turkish opium are too great, however, a potential importer could go to Hong Kong, Bangkok, Karachi, or Saigon.

Finally, as noted, trying to erect a barrier to entry by attempting to control the raw material does not seem feasible. The U.S. market relative to total world production is simply too small to prevent other supplies from becoming available. Also, the retail product is homogeneous and easily divisible, making it difficult to hoard.

However, for an illegal business, there are additional techniques of monopolization. Violence is the most likely method of securing market control; illegal operators can readily hire agents or firms willing to kill quietly and efficiently. If a heroin-distribution monopoly became established early and acquired both a large organization and a sophisticated information system, it could make it extremely dangerous for a rival to compete. First of all, it might have the key advantage of knowing who is trying to enter the business before its own prior presence can be discovered. Since rivals would find it difficult to purchase a comparable intelligence service secretly, this acts as an important barrier to entry. Another way of enforcing monopoly of an illegal product is to control the use of legal force-corrupting police officials to serve one's own ends. An effective use of corrupt police would be to have one's rivals eliminated via arrest; sufficient arrests would satisfy the public and allow the police to refrain from attacking the monopoly itself. It is in the interest of the police, in turn, to eliminate competition, protecting the source of their payoff money while appearing to be effective law enforcers. The symbiosis between criminal enterprise and elements in law enforcement is well known and has occupied many investigatory commissions without being eliminated."

A firm could also obtain police cooperation without police corruption. If one organization has better intelligence than any other, and better intelligence than the police, it could simply inform the police of the competing ring and possibly provide the evidence necessary to break it up.

The BNDD believes that 10 or 12 large wholesale operations exist in the United States today. Thus, the potential barriers to entry have apparently not been strong enough to allow any one firm to enforce a monopoly. The system has probably developed into an oligopoly or cartel, with each firm maintaining a type of monopoly control of firms below it in the lower distribution levels. It is not clear what agreements, if any, exist among these firms with respect to prices or market areas, but the risk of detection may limit the mobility of distributors between top-level suppliers, thus allowing firms at the top to maintain vertically integrated supply systems.

At lower levels in the system, there are many competitive suppliers, strict monopoly being impossible because of the difficulty of disciplining customers. At these levels, most customers are addicts who are highly motivated to maintain relations with as many suppliers as possible. Low level suppliers are arrested much more frequently than those in higher levels, because they deal more often and more openly. Thus, relationships between consumers and suppliers are constantly breaking up and beginning again. Furthermore, addicts are always trying to break into the lower levels of the system, and, given the small amount of trust, skill, and capital needed, personnel and organizations change rapidly.

Nevertheless, some elements of monopoly control remain. Information is often unavailable to consumers, thus restricting the number of markets in which they can buy. Also, addicts find it useful to restrict their purchases to dealers who consistently provide a high-quality product and who give them lower prices as loyal and trusted clients. Thus, limited information and product differentiation maintain elements of monopoly advantage even for suppliers at low levels.

COSTS AND PROFITS

Existing analysis of the economics of each segment of the heroin industry is fragmentary because of the nature of the business. The best available analysis is again that of Moore, and most of the comments in this section are derived from his model of the system.

For the importer and kilo-connection at the higher levels of the distribution system, the operating costs-principally wages and inventory finance costs-are estimated to be a much greater percentage of value added than at lower levels (50-55 per cent versus 30 per cent). Importers need skilled employees for the smuggling process and periodically require large amounts of capital to finance transactions with producers and processors. Kilo connections have relatively large wage costs because of the elaborate organization needed for exchanges and maintaining inventories as well as disciplining lower-level suppliers. Financing costs are also high at this level because of the need to store inventories for uncertain periods of time.

Operating and finance costs are more moderate for connections. Although they need to employ agents to develop new markets, discipline customers, dilute heroin, and make deliveries, there is a much faster turnover of inventories at this level. The rate of turnover minimizes financing, but connections may be forced to accept it from sources higher in the organization. Weight dealers, street dealers, and jugglers are, for the most part, self-employed; inventories are turned over quickly, and individuals at these levels rarely have financing problems.

In contrast to operating costs, expenditures for bribes as a per cent of value added appear to increase as one moves from high to low levels (2 per cent versus 5 per cent). At the importer and kilo-connection levels, well-placed bribes can secure considerable immunities at relatively low costs. Expenses for bribery increase at the connection level, because they are more vulnerable to arrest, engaging in a relatively more visible activity. They must also frequently make payments to other criminal organizations for "protection." Weight dealers, street dealers, and jugglers become even more vulnerable to extortion and will also have to pay rather large bribes (as a per cent of value added) to law-enforcement agencies.

Profits are great at each level of the distribution network but decrease as a per cent of value added at lower levels of distribution. Gross profits include (1) a financial return equal to what could be made in the firm's next best alternative line of business (its opportunity cost), (2) a return to compensate for the risk of arrest and incarceration, and (3) a return to monopoly power.

The kilo-connection at the top level of distribution is probably the single most powerful component in the system. At this level, gross profits are high and are maintained by a large degree of monopoly power. Economic theory suggests that the firm's profits must be at least as high as the return from other attractive illegal opportunities, such as loansharking or gambling. While risk-the probability of arrest-is relatively low, the combination of monopoly power and high opportunity costs makes profits at this level high. For the importer, profits are probably less than those of kilo-connections. His risks are low, because, being high on the ladder, he is difficulty to identify. Opportunity costs are low, because, with the substantial investment in building his information network, an importer can make more money dealing in heroin than he could make elsewhere. Some monopoly power exists because the personal-relations network operates as a barrier to entry, but this is probably limited by the encouragement by kilo-connections of free-lance importers.

Gross profits of connections are less than those of the two upper levels. The restricted access to information about operations allows some degree of monopolization, but risks become greater as visibility increases and the police become more successful at arrest and conviction. At this level, opportunity costs are low because of the large investment in personal relations, which are not valuable unless the connection is engaged in the heroin trade. Connections can accept lower profits, and still find the business more profitable than alternatives.

The remaining levels are characterized by relatively high risk and low opportunity costs. All three levels of operation are -relatively visible, have relatively high arrest rates, and thus require adequate compensation for risks incurred. On the other hand, opportunity costs are low because few operators on this level have the skills required in other types of enterprise. At the weight dealer level, a small degree of monopoly is maintained, because access to information is limited. At the street-dealer and juggler levels, monopoly profits are minimal.

Analyzing operating costs and profits at each stage of heroin distribution carries important implications for law-enforcement policy. In summary, if a firm's costs are high, and it has little monopoly power or attractive income-producing alternatives, it may be quite easy to increase the risk enough so that the firm leaves the industry. If these conditions do not hold-if profits are high and protected by monopoly, or if few good alternative profit opportunities exist-the firm might well absorb the cost of the increased risk. These factors largely determine the outcome of possible law-enforcement strategies.

EFFECTS OF -POLICE -PRESSURE

Increased police pressure on the distribution system can clearly increase the risks' and operating costs of those involved. Evidence does not show, however, that any one firm is large enough or monopolistic enough so that its destruction or neutralization would have an important impact on the total distribution system and thus on heroin use at the street level. During the last few years, the arrest of even major distributors and traffickers has had no significant impact on the price or availability of heroin. On the international level, for instance, in the first six months of 1971, BNDD participated in the arrest of 113 major traffickers outside the United States, seizing large quantities of drugs in the process. Federal officials face an almost impossible task of trying to stop the smuggling of drugs at the border. About 250 million people enter the country every year. Approximately 65 million cars and trucks, 306,000 planes, and 157,000 ships entered in 1970." If this were not enough, customs officials state that, on an average-size ship arriving at the Port of New York, there are 30,000 places where heroin can be hidden. Domestically, the situation is not much different. A study of the undercover unit of the New York City police narcotics division revealed that, in 1970, agents made 7,266 buys of narcotics that resulted in 4,007 arrests. However, only 4.97 pounds of highly adulterated heroin were seized in these actions, at a total cost of $91,197, and with no appreciable effect on heroin distribution citywide.

In such a situation, law enforcement must rely on the indirect effects of general deterrence rather than on direct disruption. But it is far from clear how firms operating in the field regard a successful attack on another firm. Possibly, such an attack will cause them to reassess their own risks and assume that they are in greater danger than they believed before. On the other hand, it may be that existing firms will simply assume that the destroyed operation was incompetent or lacking in security and see no need to reassess their own risk.

To the extent that police are successful in increasing a distributor's probability of arrest, one would expect the chain to lengthen as each supplier attempts to deal with fewer buyers and places more distance between himself and the street. Suppliers would invest more time and absorb greater operating costs in designing more complex delivery systems to better conceal their activities. The risk involved in holding inventories would cause suppliers to hold less in reserve, thus increasing the time and complications involved in arranging purchases and sales. However, police pressure might also increase a monopoly's hold on the system. A monopoly can absorb the increased risks more easily and is also likely to be better protected. Markets might become even more centralized. Because of higher costs and lack of competition, prices could be expected to rise.

For all these reasons, it seems clear that law enforcement can raise the price of heroin. The effects of higher prices, and thus the ultimate impact of law enforcement, are uncertain. They depend heavily on how total demand reacts to the changes.

One hypothesis is that demand is inelastic, that an addict is dependent on heroin and will try to maintain his level of consumption regardless of price. If this is so, then law-enforcement policy will have little effect. The added costs of greater risks will be shifted to consumers, and the revenues of the industry will increase. So will the social costs of the activities addicts engage in to support the habit. (A logical question is why firms are not already maximizing profits-that is, if prices and revenues could be raised before the increased cost was incurred, why did this not, in fact, take place? This may be due to concern with public reaction. Higher heroin prices cause higher crime rates, which increase the public's demand for strong policy measures against the heroin traffic. Therefore, firms may attempt to maximize profits, given the -constraint of maintaining a low level of risk. This calls for keeping prices low enough so that the crime needed to support it remains "tolerable." Competition might also play a part.)

A second theory is that demand is elastic, that a price rise does cause a disproportionate decrease in consumption. If so, then the suppliers would have to absorb the increased cost of higher risks or raise the price and accept a decline in total revenues.

While it is commonly believed that the addictive nature of heroin causes the demand for it to be inelastic, it might, in fact, be elastic, for two reasons: First, the development of methadone during the past few years has created a low-cost alternative for addicts that did not exist before. Secondly, over the last thirty years there has been a substantial long-term rise in the price of heroin relative to the price of other consumer goods. This may have produced a situation in which most addicts are maintaining their habit at the highest price level they are able to support. If the price increased further, these addicts would have no choice except to decrease the quantity of heroin used, because they would be unable to raise their income enough to meet the new cost.

If the demand is elastic, a law-enforcement policy could be effective. If the price of heroin increased and methadone was necessarily substituted, the social costs of addiction, including crime, would decrease. If addicts chose to adapt by using less heroin for the same cost, however, the crime costs would not be decreased. It could be argued, on the other hand, that increasing the cost of being an addict hastens the time when the addict voluntarily leaves the addicted life.

The firm attempting to maximize profits, subject to the constraint of maintaining a low level of risk, may choose to absorb the risk as increased operating costs rather than attempt to shift it to the consumers. This, of course, results in a lower rate of return and reduces the attractiveness of heroin distribution. In the heroin business, the gross profits of a firm (revenues minus operating costs) must be at least equal to the return the firm could earn in an alternative investment (its opportunity cost).

Profits that exceed opportunity costs can be considered a return to a monopoly position or compensation for risk. Gross profits must exceed opportunity costs by at least enough to compensate the firm for the risks it undertakes. If gross profits, less opportunity costs, exceed the compensation required for risk, the firm will stay in the business. Alternatively, if gross profits, less opportunity costs, are less than the return required for undertaking risk, the firm will not be willing to absorb the risk and will leave the industry.

Risk for an illegal firm can be viewed as equal to the probability of punishment, as the firm perceives it, times the cost of certain punishment to that firm. The cost of certain punishment is equal to the amount the firm would be willing to pay to avoid it.26

The probability of punishment can be affected by public policy and by the firm's own actions, because it depends on the probability of being apprehended and successfully prosecuted. Changes in the criminal-justice system are essential if high-level distributors are going to be put out of business. These include more sophisticated training of police personnel, higher salaries to lessen the temptation of bribes, and higher budgets for purchasing information. In reality, law-enforcement agencies have been unable to raise the level of risk anywhere near what would be required to make this criminal enterprise unprofitable.

The cost of the expected punishment to the illegal operation can also be influenced by changes in judicial procedures and increases of sentences. Currently, many defendants in narcotics cases forfeit bail, become fugitives, and, if apprehended, are more likely to be tried for violating the conditions of bail than for the original charge. If convicted, this means a lesser sentence. The economic cost in such a case-the amount the offender would pay to avoid detection-should be equal to the amount of bail, plus the probability of being arrested while. at large, times the penalty for jumping bail. This cost would be increased if a change in public policy raised the amount of bail or the penalty for violating its conditions .21

On the other hand, the probability of punishment can be reduced if the illegal operation spends more on bribes and devising elaborate, less conspicuous pickup and delivery systems. If increased amounts are spent to reduce risks, however, the illegal firms reduce their gross profits at the same time. A firm will expend these sums if in doing so it can keep its gross profits (minus opportunity costs) above the 'cost of escaping imprisonment. If it is impossible, through further spending, to reduce the expected risk below gross profits, less opportunity costs, firms will leave the business. Alternatively, firms could limit sales to trusted buyers, thereby reducing revenues and gross profits but also reducing the chances of getting caught.

According to Moore's analysis of the distribution hierarchy, gross profits are considerable at every level. However, for the kilo-connection, high-yielding alternatives are the greatest, and it is possible that monopoly at this level could be broken. BNDD appears to believe that the monopoly has already been broken, and that the industry is now operating as an oligopoly with ten or twelve firms. But, conceivably, an oligopoly or cartel could develop techniques for supplying the total market as efficiently as a centralized organization. If satisfactory agreements can be reached, the amount of explicit communication necessary to guarantee avoidance of frequent shortages or surpluses would be minimal. However, the market becomes more unstable with a cartel, because there is always an incentive for one firm to increase its sales at the expense of another. If one firm does try to improve its position, the others are likely to follow. If all try to do so, oversupply develops, attempts to expand the market increase, prices decline, and profits fall. In the process, higher levels of the distribution system become more visible to the authorities, and the probability of punishment increases. A kilo-connection may then find one of his high-yielding alternative opportunities more attractive and leave heroin distribution. Thus, there is always some incentive for the police to attempt to infiltrate the network at higher levels. Arrests made at this level reverberate through the system, making arrests easier at all levels.

It is far from clear, however, that such a public policy is altogether sound. While it could have a high payoff in terms of seizures of heroin and arrests, it would lead to considerable market expansion and increased heroin availability. The elimination of monopoly would permit the entrance of new firms; a competitively organized industry will always provide more goods at lower prices than will a monopoly. Market expansion would occur if lower prices increase the demand for heroin, and, in fact, this may be occurring. Two very large seizures by BNDD in May and June of 1970 reportedly had no effect on price. No panics occurred following the seizures, as one would expect, and the short-term trend over the year was one of declining prices.

Theoretically, the best allocation of police power is at the levels of the distribution hierarchy, where the incremental gains from their actions (in terms of reduced social costs) are equal to the incremental costs of the enforcement efforts. This clearly requires employment of more resources at some levels than at others. While the payoff from attempts to infiltrate higher levels may be large, this policy is also very expensive. And, again, it is possible that such a policy would lead to a larger, less organized trade. Use of police resources at lower levels has less potential for large payoffs but is less costly. It is much easier to interrupt the heroin traffic at the retail level, because firms engage in more transactions, concealment is more difficult, and information is more available.

On the other hand, risks must be increased considerably before it becomes no longer profitable for low-level operators to remain in the industry, because their opportunity costs are low-i.e., they probably could not earn nearly so much doing something else. Police are continually eliminating these low-level operators through arrest, but, given the high profit potential, others still find it worthwhile to enter the business because of their low opportunity costs, mistaken perceptions of the probability of punishment, or the low value placed on being incarcerated for a given period. In addition, most street dealers and pushers are themselves addicts. Being in the business keeps them near the source of supply, giving them an income sufficient to finance their own consumption and enabling them to buy their own drugs wholesale.

Clearly, if heroin demand becomes relatively elastic at some point as prices rise, the police could eventually be successful in eliminating the industry. However, to accomplish this, it is probable that society would have to allocate an extraordinary amount of law-enforcement resources. Society might prefer to reduce the industry to the "optimum" size, where the costs imposed by heroin use are just equal to the additional cost required to reduce it further.

CONSUMPTION

The consumption of heroin imposes costs on the rest of the community, because most purchases are financed through criminal acts such as theft. Because of these involuntary redistributions of existing wealth, the community spends resources to prevent crime and to apprehend, try, punish, and rehabilitate criminals. This includes private expenditures on locks, alarms, lights, and security guards and public expenditures on police, courts, and correctional systems. To the extent that these human and material resources would be employed elsewhere in the absence of heroin addiction, they are properly counted as costs of addiction. Social costs also include the medical expenses and foregone productivity of individuals injured in crimes and the lost productivity, medical expenses, and premature deaths of addicts. Finally, one must also include the fear and anxiety, avoidance of normal activity, disruption of community life, and use of less efficient and more expensive modes of transportation that are unquantifiable but nonetheless very real costs of crime and drug addiction. This section examines the direct cost of heroin consumption and estimates the amount of crime required to finance it.

The most elementary question in determining the direct costs of heroin consumption is the total cost of the heroin consumed that is, how much money do addicts need to support their addiction? But this calculation depends on several unknowns: the number of addicts, the average number of days per year heroin is consumed, and the average cost of heroin per day, with an adjustment for the amount of heroin "earned" from profits on sales to other addicts." This information is essential for determining the magnitude of the problem and the benefits from any investments designed to control or reduce it.

THE ADDICT POPULATION

The actual number of heroin addicts in the United States is unknown and probably can never be determined precisely with present techniques of estimation. Addicts become known to officials only through death, arrest, or participation in treatment programs. It is difficult to estimate the size of the addict population from the number of deaths or arrests, because each of these figures depends on other variables as well as sum totals. For example, the ratio of addict deaths to addict population will vary not only with the number of addicts but also with the probability that a decedent will be medically identified as an addict. Similarly, the frequency of addict deaths from overdose will vary with sudden changes in the concentration of pure heroin in the local supply. The ratio of addict arrests to number of addicts can vary with the intensity of police efforts in certain areas. The ratio of addicts in treatment to total addicts can vary with the addicts' perception of the value of treatment or the short-term availability of heroin.

Because of these difficulties, most estimates of the addict population are mere guesses. However, there are two methods that may ultimately prove useful, although both have pronounced weaknesses at present. One is the "Baden" formula, developed in New York City by Dr. Michael Baden of the Medical Examiner's Office, and frequently applied elsewhere. New York is unique in having a register of local heroin addicts as reported by medical and police sources. The formula was computed by comparing the number of addicts on the registry to the number of overdose deaths and the number of deaths among persons named on the register." Baden found that 50 per cent of the heroin addicts who die are on the Narcotic Addict Register. On this basis, he estimated that the total addict population numbers about twice the registry total. In 1968, for example, the New York Register had 42,500 addicts listed; it was therefore concluded that the total addict population was approximately 85,000.

This method has several problems, of which its originators were well aware and are quick to acknowledge. They were seeking a way of making a quick estimate of the total population that would be more adequate than the register alone without incurring expenses for research time. (It is ironic that Dr. Baden is sometimes castigated for defects that he was the first to point out.) First, it is assumed that addicts who are on the register have the same probability of dying as those who are not. If registered addicts are heavier users and thus more susceptible to disease, such as malnutrition and hepatitis, they would have a higher probability of dying. On the other hand, if addicts not on the register are less experienced users they may be more likely to die of overdose. Either, both, or neither hypothesis could be correct, and it is impossible to determine the net effect.

Secondly, since the register is cumulative over a five-year period, some on the list will no longer be addicts because of successful treatment, abstinence, or incarceration. To the extent that this is the case, the estimate will exceed the actual total by twice the size of the error in the register.

Third, the probability that an addict is listed on the register is, in part, a function of his age and length of addiction. A thirty year-old addicted for ten years is more likely to be listed than an eighteen-year-old addicted for one year. A sophisticated application of the technique requires at least an age breakdown, and possibly other distinctions as well. Baden himself works with such distinctions, but others who use the method often do not. Serious difficulties develop when officials attempt to use the Baden formula in other cities or at other points in time. Statistical analysis of one set of circumstances is rarely transferable to different situations. In other words, heroin addicts in different places, for whatever reasons, probably do not die at the same rate. Furthermore, differences in diagnostic method change the probability of a decedent's being identified as an addict. For example, in Washington, D.C., during the eighteen months preceding July, 1970, diagnostic drug screens were performed on only 6.3 per cent of all autopsied deaths; during the last six months of 1970, drug screens were performed on 51 per cent of all autopsied deaths." This frequency can also vary between cities, and a small variance, when using Baden's formula, would change the estimated population greatly, because the number of addicts who die in any given year is likely to be only 1-2 per cent of the total register." And New York has by far the best register of any city.

The Bureau of Narcotics and Dangerous Drugs has developed a somewhat similar approach in estimating the total addict population. It multiplies the number of known users who could be rearrested by the ratio of total heroin arrests to the number of known users included in these arrests." On the basis of this method, BNDD unofficially estimates that there are at present 315,000 addicts nationally.

This method also has several problems. First, known addicts are voluntarily reported to BNDD by local police departments. Reporting practices can vary considerably with locality, as can intensity of enforcement efforts.

Secondly, the probability of any addict's being arrested in period 2 is assumed to be the same regardless of whether he had been arrested in period 1. This hypothesis is unproved. It may be that addicts already known to the police are heavy users and commit more crimes or are under greater surveillance. Therefore, the probability of their rearrest would be greater, and the estimate of the total addict population would be less than the actual. On the other hand, if the addicts unknown to the police are younger, and therefore more recently addicted, the probability of their arrest may be higher, because younger addicts tend to commit riskier crimes, such as robbery. In this case, the estimate would be too large."

Thirdly, it is difficult to determine the number of addicts reported in period I who remain active users and are therefore eligible for arrest during period 2. Statistical adjustments must be made for the probability that a known user will be in jail during period 2, will die, will become abstinent spontaneously, or will be in a treatment program. BNDD has thus far adjusted satisfactorily only for the probability that an addict will be in jail.

Some of the problems involved in these adjustments are shown by the difficulty of estimating the probability of death. BNDD made its adjustments on the average probability that any individual over twenty-five will survive to the next year. But the mortality rate for addicts is undoubtedly higher than that of other individuals of the same age. On the other hand, there may not be many addicts over forty, and thus the mortality rate as originally calculated was biased upward by calculating an average probability for all individuals over twenty-five. Without knowing the true mortality rate, it is difficult to make the proper adjustments to the estimates.

Some addicts do mature out of addiction as they age, but the number is not known. Even active addicts may have weeks or months of abstinence, but again no numbers are available. Work on these problems is going on, and the results will probably reduce the estimate below the present 315,000. However, since reports from various urban areas indicate that the number of addicts is still increasing, we will use alternative estimates of 250,000 and 350,000 for the total addict population.

Some officials make a "best guess" estimate and then attempt to test the reasonableness of the figure by comparing it with the total population. For example, the Narcotic Treatment Agency in Washington, D.C., has a population of 95 per cent black, 85 per cent male, and 85 per cent between the ages of fifteen and thirty four. Assuming that the addict population not in treatment has the same characteristics, one can make an estimate of the penetration of addiction into the total population with these characteristics. If there are 15,000 addicts in Washington, 10,296 are black males between the ages of fifteen and thirty-four. According to 1970 Census data, there are a total of 83,158 black males between fifteen and thirty-four in Washington, D.C. Thus, 12.1 per cent of this population is addicted under the assumption of 15,000 addicts, and 8.2 per cent is addicted under the assumption of 10,000 addicts. The problem with these projections is that 8.2 per cent and 12.1 per cent seem equally 'reasonable" If an estimate approached 25 per cent or 30 per cent of a given population, one would suspect exaggeration. Penetration projections appear to have little value except as very rough assessments.

DAILY EXPENDITURES FOR HEROIN

Addicts vary in the amounts of heroin they consume per day and the per cent of time spent actively using the drug. Some addicts can afford to increase their habits as their tolerance to the drug grows; others cannot. There is also great variance in vulnerability to arrest and incarceration as well as in the frequency of "vacations" from addiction in hospitals or treatment facilities. These periods of treatment also serve to reduce the size using interviews with ex-addicts and police, developed a breakdown of the addict population of New York City." From their data, presented in Table 4-5, one can estimate that an average addict is an active user 70 per cent of the time, or 255 days per year per addict.

The frequently quoted estimates of the daily costs of supporting an addiction are often exaggerated. Many are derived from interviews with addicts who regard large habits as status symbols and commonly overstate the level of their own use. In general, the amounts of heroin consumed by addicts vary widely, and reports of these amounts are unreliable. To say that an addict has a $50 habit does not mean that he will spend this sum day after day. Ideally, an addict would like to use as much heroin as he needs to "get high." In fact, however, the increasing difficulty of obtaining the necessary money to pay for increasing dosages imposes an upper limit on most addicts' habits. But if an addict cannot consistently obtain large enough dosages of heroin -to get high, he can use smaller amounts without feeling acute withdrawal symptoms. In effect, the habit size of a long-time heavy user is often not at the level of euphoria but only at a level sufficient to suppress withdrawal or to keep the withdrawal mild. Moreover, while there are many addicts with expensive habits, there are also many addicts in the early stages of dependence whose habit requires only $10-15 per day. These factors should be kept in mind when interpreting any estimates of the total cost of addiction.

The price, and quality, of heroin needed by an addict determine his cost per day. Retail prices were obtained by BNDD -for December, 1970, for several U.S. cities, and are listed in Table 4-6. The considerable variation in price is probably due to differences in local supply and demand, risks taken by suppliers, and, probably, nonrandom buying procedures. The unweighted average price-except in Chicago, where it is extraordinarily high for unknown reasons-is about $0.55 per milligram.

The average quantity used per day is probably about 55 milligrams of pure heroin." At $0.55 per milligram, this results in daily expenditures of approximately $30 per day for the average addict, representing the amount paid on the street by the addict who buys solely for his own consumption. Some earn their heroin through services rendered as a distributor of heroin-i.e., pusher; thus, the price paid by the ultimate consumer includes the payment for that service. Heroin "earned" in pushing, therefore, means that fewer addicts have to steal from the community, or that some percentage of the addict population supports the rest by theft, prostitution, or legitimate work. The Hudson Institute study estimated that 45 per cent of the heroin consumed is used by those involved in its distribution and profit from its sale. If other factors remain the same, the total social cost of heroin addiction will be less if the percentage of consumed heroin "earned" from pushing is increased.

A rough estimate of the annual cost of heroin purchases can be made using the information presented in the preceding pages. The assumptions are that (1) addicts spend 70 per cent of the

.

SOURCES OF FUNDS FOR HEROIN PURCHASES

The high cost of heroin requires most addicts to resort to crime to finance their habits. Estimates of the various sources of funds for the support of heroin addiction can be made with data obtained by the Hudson Institute from ex-addicts in New York City. As discussed previously, approximately 45 per cent of the heroin consumed by the addict population is obtained by selling heroin. This does not mean that 45 per cent of all addicts are pushers but that 45 per cent of the heroin used by addicts is obtained in exchange for services rendered in the distribution network. The remainder of the addict population finance their own habits and those of pushers by various means, including robbery, burglary, shoplifting, prostitution, and legitimate work. Most addicts employ a combination of means, legal or otherwise, in their constant quest to obtain funds." The breakdown of sources of funds presented in Table 4-7 was developed from the Hudson material. These figures represent the funds used to purchase the 55 per cent of heroin consumed that is not financed by rendering services in the distribution network.

It is assumed that shoplifting, burglary, and larceny yield real property, and that pickpocketing, robbery, and confidence games yield cash. Therefore, the total property stolen in these crimes can be roughly divided into 80 per cent real property and 20 per cent cash." It is assumed that real property is fenced at one-third its value, so that an addict must steal, on the average, property worth $2.60 to obtain $ 1.00 with which to finance his habit."

The data needed to estimate the funds obtained through property crime, prostitution, and legal sources used for heroin purchases can be summarized as follows:

Number of addicts (N) 250,000; 350,000

Per cent of year in active use (T) 70

Average daily expenditure (X) $30

Per cent of heroin earned through pushing (d) 45

Proportion of funds obtained through property crime (a) .62

Reciprocal of fencing discount (average) (f) 2.6

Proportion of funds obtained through prostitution (b) .31

Proportion of funds obtained through legitimate sources (w) .07

The estimates of funds obtained from each of the three alternative sources of funds are presented in Table 4-8. Clearly, any of these data could be biased in either direction and thus limit the usefulness of the estimates derived from them.

CONCLUSION

The economics of heroin distribution has been examined for whatever assistance such analysis may bring to future policy considerations. Can heroin use be curbed? Can it be eliminated? What are the most efficient methods of attacking the problem?

In the marketing of any product, there are two sides to a transaction-supply and demand. Our examination has shown that public policies dealing with the production of opium and the domestic distribution of heroin-the supply side-will never be completely successful in eliminating heroin use in the United States and the consequent street crime. Indeed, past reliance on law-enforcement measures has accompanied an epidemic growth of the heroin-addict population. It is ironic that our national policy, by resorting almost exclusively to criminal sanctions to eliminate addiction, has bred a thriving illegal market that can be sustained only by criminal activity.

Two factors primarily impede public policy designed to suppress the supply of heroin. First, the amount of opium needed to supply the demand for heroin in the United States is only a minute portion of total world production. It is probably impossible to curb all opium production, licit and illicit, everywhere in the world. Secondly, the profits earned in domestic distribution are so great that it is unlikely the risks can be raised high enough to force dealers out of business. As one former distributor has stated: "If you put me away, somebody else is going to stand in my spot.""

Clearly, more effort and greater resources must be spent to curb the demand for heroin. For example, more addicts must be treated and rehabilitated. Education to prevent drug abuse must be expanded and made more meaningful. Finally, the environmental factors that breed and sustain addiction must be more successfully attacked. John Ingersoll, Director of the Bureau of Narcotics and Dangerous Drugs, has stated before a congressional committee that law enforcement is the first-aid agency of society. It is not the curative, not the doctor, and it doesn't eliminate causes of these problems. Our society has gotten itself into an unfortunate state of affairs regarding drugs because it has assumed that by passing laws and enforcing them, the problem will go away.

So it seems to me that while you can expect law-enforcement agencies to do the first-aid work, if you want a cure, then we have got to go back to the basic causes and find out what can be done from that end. We have to look to improved and increased rehabilitation programs. We have to find out more about drugs and why people use them, knowing of the debilitating consequences .42

No policy, of course, will. be successful if it addresses only a part of the problem. There appears to be, for instance, a causal relationship between strict law enforcement and the number of addicts in treatment programs. If the probability of arrest is increased, more addicts are induced to enter treatment. Conversely, if more hard-core addicts enter treatment programs, pushers must seek out new, less reliable clientele; they become more visible and thus more vulnerable to arrest. Development of an effective policy combining efforts to suppress supply and inhibit demand for heroin necessarily involves close scrutiny of the economic system and the relative benefits and costs of various alternatives. In choosing between policy alternatives, cost-benefit analyses can illuminate the probable reduction in the social costs of addiction to be derived from one policy as compared with another. If our public monies are to be spent efficiently, it is imperative that such choices be illuminated by full-scale studies of this nature.

NOTES

1. BNDD Staff. Interviews, November, 1971. This assessment is based on two other estimates: the total number of addicts in the United States, and the quantity of pure heroin consumed by each addict per day. The accuracy of these estimates will be analyzed later. But just one year ago, BNDD believed that only 5,000-6,000 lbs. of imported heroin was sufficient to supply the U.S. market. The current estimate should be interpreted accordingly; the actual total is probably somewhat less than 10,000 lbs.

2. BNDD Intelligence Staff, The World Opium Situation (unpublished paper), October, 1970, p. 13 (hereafter cited as WOS). This assumes that the illicit production estimates are accurate,

3. Ibid., p. 16.

4. Select Committee on Crime, Hearings into Narcotics Research, Rehabilitation and Treatment, U.S. House of Representatives, 92d Cong., 1st sess., Serial 92-1, 1971, p. 358.

5. WOS) p. 6.

6. Ibid.

7. Ibid., p. 4.

8. International Narcotics Control Board, "Statistics on Narcotics Drugs for 1967." International Narcotics Control Board (New York: United Nations, 1970), p. viii.

9. WOS, p. 5. This publication is the chief source of the data in this section.

10. It is extremely difficult to estimate illicit production and consumption. Except for Turkey, illicit production estimates were made on the basis of domestic consumption. Some allowance for exports is included but is probably subject to error. The BNDD staff considers the estimates to be accurate within a confidence interval of 20 per cent.

11. BNDD Staff, interview, March, 1971. Indian opium is often adulterated with seed, leaves, and other foreign matter. Indian yields are also higher because of additional lancing.

12. United Nations Survey Team, "The Hill Tribes of Thailand and the Place of Opium in Their Social Economic Setting," Bulletin of Narcotics, XX, No. 3 (July-September, 1968).

13. WOS, p. 7.

14. G. Shulzin, "Cultivation of the Opium Poppy and the Oil Poppy in

the Soviet Union," Bulletin of Narcotics, XXI, No. 4 (October-December, 1969).

15. BNDD Staff, interview, March, 197 1.

16. Select Committee on Crime, op. cit., note 4, p. 68.

17. These figures are approximate. In fact, the necessary land area could be considerably less, depending on the number of total addicts and the opium yield per acre. The 11.2-23.2 square-mile estimate is obtained from the data in Table 4-2. For example, if the yield per acre is assumed to be 8.07 Ibs. (the approximate reported yield in Turkey), then the area required to produce 100,000 and 120,000 lbs. of opium is approximately 19.4 square miles and 23.2 square miles, respectively. (If 1 acre yields 8.07 lbs., then 12,391.5 acres will yield 100,000 lbs. This acreage is approximately 19.4 square miles.) On the other hand, if the true Turkish yield, as suspected, is approximately 14 Ibs. per acre, then the area required to produce 100,000 and 120,000 lbs. is approximately 11.2 and 13.4 square miles, respectively.

18. WOS, p. 32.

19. WOS, P. 8.

20. If demand is price-elastic, the percentage reduction in quantity sold exceeds the percentage increase in price. thus reducing farm revenues.

21. Mark Moore, Policy Concerning Drug Abuse in New York State, Vol. III: Economics of Heroin Distribution (Croton-on-Hudson, N.Y.: Hudson Institute, 1970), pp. 67-73.

22. John Casey and Edward Preble, "Taking Care of Business: The Heroin User's Life on the Street," International journal of the Addictions, IV, No. I (March, 1967),12,

23. T. S. Schelling, "Economics and Criminal Enterprise," The Public Interest, No. 7 (Spring, 1967), 61-78.

24. Ibid.

25. Select Committee, op. cit., p. 3.

26. R=pV, where R is the cost of risk to the firm, p is the probability of punishment as perceived by the firm, and V is the value, or cost, of the expected punishment to the firm.

27. Select Committee on Crime, Hearings into Crime in America: Heroin Importation, Distribution, Packaging and Distribution, U.S. House of Representatives, 91st Cong., 1970, pp. 52-91.

28. The annual cost of heroin consumption can be stated by this equation:

CH=365 [N (1-d) TX], where
CH=net annual expenditures for heroin;

N =number of heroin addicts;

d =the percentage of all heroin consumed by the addict population, which is obtained for services rendered in the distribution network-i.e., pushing;

* =average percentage of the year heroin is consumed by addicts; and

* =average cost of drugs per day to the addict population.

29. Baden's formula is:

N = D - n, where

d

A

N=estimated total addict population;

D =total number of deaths of addicts in given year; n = number of addicts on Narcotic Register in given year; d=number of deaths of addicts on Narcotic Register in given year; and N =actual total addict population.

30. Robert J. DuPont, "Profile of a Heroin Addiction Epidemic," in Select Committee on Crime, op. cit., pp. 4,167.

31. For example, the derived n in Baden's formula for New York City

d

has been used to estimate the number of addicts in Washington, D.C. The equation becomes:

n

N D, where

A

N =estimated total addict population in D.C.;

n =number of addicts on New York City Narcotics Register; d= number of deaths of addicts on New York City Narcotics Register in given year; and

* =total number of addict deaths in Washington, D.C.; and

* =actual total addict population in Washington, D.C.

The problem here is that differences in capabilities for identifying narcotics

n n n

overdose deaths can make 7 vary among cities. If d in DC > d in New

n

York City, then N < N; alternatively, if in D.C. < n in New York City,

d d

then N > N. Clearly, similar problems arise with efforts to use n for time

d

period I to estimate the addict population in period 2 with the total number of deaths in period 2.

32. BNDD Staff. Interviews. Included are only those important features of

the method necessary for clear description. Their equation is:

A t

T = N - , where

n

A

T =the estimated total addict population in period 2;
N =the number registered with BNDD in period 1, who could be arrested in period 2;

t =the total number who are arrested in period 2; n=the number registered with BNDD in period 1, who are arrested in period 2; and

T =the actual number of addicts in period 2.

Once it is known how many addicts have been arrested in period 2 (t), and what the number of these previously registered with BNDD in period 1 (n) is, then N can be used to estimate T.

33. As discussed in the text, the BNDD method assumes n = t But, if m

N T.

is the number of addicts arrested in 2 who were not arrested in 1, and if M is the total number of addicts not registered with BNDD in period 1, then

n . m must hold for the method to yield true estimates. If n > m then

N M N M

A

T < T; similarly, if n < m then @ > T.

N M

If n = M a particularly useful feature of this method would be the fact N Mthat new entrants in the current period do not bias the estimate of the total population.

34. Moore, op. cit., note 21, p. 66. These estimates were made on the basis of a relatively small number of interviews and should not be interpreted strictly. However, this table provides the best available information on these aspects of addict behavior.

35. This estimate, while lower than several others, may still be too high. First, the Hudson Institute average of 55 milligrams includes an estimate that large-habit dealers (I 1 per cent of the total user population) use 180 milligrams per day. While heroin dealers might have the income to maintain such a large habit, it seems unlikely that they could be active entrepreneurs and at the same time use such quantities of the drug. The Hudson estimate was also derived from interviews with addicts, who reportedly exaggerate estimates of use. So, too, there may be less than 10 mg. of pure heroin per bag. 36. The amount of funds obtained from each of three alternative sources of funds (property crime, prostitution, legal sources) can be estimated with the following equation:

C=365 [ U -d) NTX (af +b+w) where d, N, T, and X have the same meaning as in the equation in footnote 28.

C --the gross -amount uf -funds ubtained -annually -from various -sources -for heroin purchases;

a=the proportion of funds spent on heroin that comes from property crime;

b= the proportion of funds spent on heroin that comes from prostitution; c=the proportion of funds spent on heroin that comes from legitimate sources-i.e., work, public assistance, family, or friends;
f =the factor by which the amount stolen must exceed the amount yielded by a property crime. (A stolen good must be transferred from the thief to a fence to a final consumer. The thief receives only a fraction of the market value of the good. To obtain a given target yield, an addict can be expected to steal a given amount, depending on the fencing discount, and the proportion of value stolen that is property and not cash.)
The amount of money annually spent by addicts for heroin can be estimated with the equation in footnote 26.

The cost to victims of property crime committed by addicts can be estimated by this equation:

Cv = 365 [ (I - d) NTX (af)

where C, =cost to victims of property crime.

37. These estimates are also weakened by the interview technique em-

ployed and thus should not be considered a precise reflection of addict behavior.

38. The value of property stolen cannot be counted totally when the costs to society are added. Some percentage of the value of goods stolen accrues as a type of subsidy to people who, knowingly or unknowingly, purchase stolen goods because the price they pay is significantly below the market price. This factor reduces the cost of crime to all society by the value of real property stolen multiplied by the reduced resale value (the fencing discount).

However, in ascertaining how large an adjustment to make, it is difficult to compare the utility of gains to the purchaser with the utility of the losses to the victim-is it greater, less, or the same? In addition, from the point of view of the community, the loss to the victim may be given more weight than the gain to the ultimate consumer. The item may, in fact, lose "value" because it is a stolen good. Objectively, although a stolen television set may still operate as well as before, the purchaser may feel that it is worth less than the market value because of the risk of holding a stolen good.

39. '80 +.20 ($ 1) = $2.60

1/3

40. The net annual addict income from property crime can be estimated

by the following equation:

PC= 365 [ (1 - d) NTXa]

where d, N, T, and X have the same meaning as in footnote 28 and

PC= net annual addict income from property crime.

The cost to the victims of the stolen property is determined by multiplying

the net addict income from property crimes by the reciprocal of the average

fencing discount.

41. Select Committee on Crime, op. cit., note 27, p. 238.

1. Ibid., note 4, p. 367.