Interviewing Gary Becker...

1

This conversation was held in 1995, February 26th at 14.00 p.m. In his office, 5 floor in the building of Social Sciences, University of Chicago, by Javier Sainz and Enrique Sainz, while this one was visiting professor in the University of Chicago with James Coleman. The conversation follows an more ample conversation held in Palace Hotel in Madrid at the time Gary Becker was invited by the Banco de España. Question: A. Addictive behavior and the rational choice theory. Addictive behavior is both a challenge and a canonical model for an econometric analysis of behavior according to the rational choice theory and it may represent well many of your research interests in topics like human capital with a particular emphasis in productivity and investment in skills, knowledge and education, like the construction of theories of allocation of time, of marriage, of family and demography, and like law enforcement and crime . Other topics are also connected to this important one, giving to your professional trajectory an impressive unity: How did you begin to become interested in addictive behavior? Gary Becker: I became interested in addictive behaviour, in a sense, in an accidental way. I was doing a paper with J. Stigler called De gustibus est non disputandum and I began to think about the issue along withStigler; some examples have been given in the literature when there is not diminishing marginal utility, and Alfred Marshall, a famous british economist have an statement that the interesting good music is an exemption to the law of diminishing marginal utility because the more good music a person hears the more he wants to hear and in thinking about how to incorporate that into a model of rational behavior it ocurred to me that the natural way to do that was to assume good music was a habitual addictive good, the more you have it the more you want to have it and so I started to formulate a model of that, we did it, we published that in 1977 I believe, and then I didn't do any much more on that problem for a while I kept thinking that much more could be done on addictive behavior than we did. And I had a student who wrote a thesis on addition, Larry Ianaccone, a very good thesis, and that stimulated further thinking in my part and I began to work jointly with Kevin Murphy and we produced a series of papers together and so on. Question: Your research extend the traditional theory of individual rational choice to analyze social issues beyond those usually considered by economists incorporating into the theory a much richer class of attitudes, preferences and calculations; micro level analyses are used as tools to derive implications at the group or macro level; subjects are motivated by selfishness or material gain, but this is a method of analysis, not an assumption about particular motivations. Is there any irrational behavior? Since it is argued that time preference for the present may go to infinity, irrational behavior, that is behavior that ignores future consequences of a change in current consumption, may become rational, is some place left in the rational addiction theory for a completely irrational behavior? Is not this reasoning circular, a petitio principii? Gary Becker: A natural question that people ask as well if this a model of addictive behaviour, is there any such a thing as rational addictive behaviour?, I should say, is there any such thing irrational behaviour or in particular, addictive irrational behaviour? That is a tough question, I assume there is such behaviour that would be inconsistent with our analysis, one way we... one example of that may be behaviour that is not forward looking, so future consumption will not have any influence on my present consumption. Now, that is a possibility, naturally, we try to test that in our, in a paper that is coming out on cigarettes "... demand approach to smoking#" and what we do is contrast the effects of the rational addictive approach to smoking with particular versions of irrational, non-rational and the versions we test are those which ignore the future and make decisions based only on past and the present. And our conclusion is that a model which ignore the future do badly explaining smoking in the US. The rational addiction model is not to do perfectly but it does much better and that paper will be I... try to document that point. So yes there is irrational behaviour, in particular irrational addictive behaviour, but I do not believe there is nearly as much of that as people claim there is and the people confuse irrational behaviour with other types of behaviour that are perfectly consistent with our approach, the rational behaviour. Question: If rationality of addiction cannot be an assumption of the model it must be a consequence of the way addiction is ultimately accounted for by a rational model. According to Cognitive Science for a behavior to be considered rational it should be able to be systematically changed by changes of information. Under certain circumstances subjects may have greatly reduced possibilities of processing information as under an abstinence syndrome... Gary Becker: One issue sometimes raised is rather how information changes smoking we should see a very good example of that in an experience in most countries since early sixties when more information became available on the cancer health consequences of smoking, we know in every country overall smoking is declined and in U.S is declined

Interviewing Gary Becker...

2

enourmously, and in particular is also declined not only for mature person or educated people but also decline for young people and the less educated. What is true is that more educated people have a much bigger decline that less educated people and my explanation is that educated people have better information or at least are better able to process the information than less educated people so I don't take that as an example of irrational behaviour but as some failing in the ability of any less educated people to have any realistic process of that information. On the other hand if you look at or, let me put... , I don't want to say that, I want to say more than that. There are also properly rational reasons why the less educated people should smoke more than the more educated even if non-smoking is better for your health, the value of the lives of that less educated people in terms of the utility are less, their earnings are less, what they forgo by getting sick are less, so, you know, there are many reasons, also the less educated people have more difficulties in their lives and we show, one reasons people continue smoking is because they are hit by varied shocks so I like, I think it has to be greater qualified what I said at the beginning I wouldn't ... I don't think that is a right view it may be that less educated smoke more because they can't process information as well but there are other reasons what they aren't doing as well and I am not yet sure exactly what is the right explanation. If you look at some surveys that have been done at people what they considered to be the likely of the having lung cancer from smoking, most people overstated the true probability in their analysis rather than understating, it isn't that people seem to be understating the likelihood, they always states the likelihood, so wether the smoking is caused by imperfect capacity to process information isn't completely clear to me and I don't ... know enought about that yet. Question: Rational decisions are made over time, purposive behavior is goal-directed; addictions are treated as an example of rational purposive behavior, as goal-directed behavior where the goal is maximizing utility over time. However, there is no way a priori to maximize utility over time unless actual utility is compared with past utility. The question is, do subjects try to maximize utility over time or just get as much utility as they obtained in the past? If not, under what criteria do they try to maximize utility? what role may be assigned to memory of past experiences in current behavior? Subjects may prefer to be conservative instead of looking forward at increases of utility. It has been shown that bees refuse to suck the juice of flowers with much more food if they are to be found in areas with less food on average. What relation may be traced between immediate utility and differed utility? Gary Becker: Well some issues are raised here by maximization of the utility over time versus margins of the utility from the past, I think memory of the experiences that they have in the past is relevant, we incorporated that in our analysis by assuming the effects of the past depreciate with time so you can interpret that... as a limit of memory of the past, may be different people's memories would be different and what we show is that the more rapidly the past gets depreciated the less likely peple are to be addicted so addition is posibly related, I am sorry let me put that a bit differently... the more rapidly people depreciate the past the more likely they are to become addicted, nor less likely, because that means when they smoke now it doesn't have such painful... the expected harm in the future is less because they know it is going to depreciate rapidly so that is important and that is a factor that explain that some people of poor memories they are more likely to become addicted, so I think that is relevant but that is not inconsistent with rationality. Question: B. Modeling rational addiction Basic concepts and rational theory of addiction. We can submit you a certain number of questions... "Addictive behavior is usually assumed to involve both "reinforcement" and "tolerance". Reinforcement means that greater past consumption of additive goods, such as drugs or cigarettes, increases the desire for present consumption. But tolerance cautions that the utility from a given amount of consumption is lower when past consumption is greater" (1991, p. 237). Reinforcement and tolerance seem to have additive effects on consumption: consumption must increase with tolerance to get the same utility as in the past what lead to greater consumption and reinforcement increases consumption of addictive goods over time (as long as reinforcement has a positive effect on increases of consumption capital); what mechanism is ultimately responsible for an increasing consumption of addictive substances, reinforcement, tolerance or both? could we separate in the theory the effects of both mechanisms? Addictive substances may have beneficial effects given a rate of consumption and become harmful given another rate usually greater, how this fact affects addictive capital? Why subjects would rather addictive goods than non-addictive goods to cope with stressful events? How this preference would affect calculations of utility over time? Were subjects rational they would always prefer beneficial addictions than harmful addictions, may we separate addictions from the existence of a market of addictive substances? Consideration of harmful consequences of their current behavior may never actually happen if rational utility

Interviewing Gary Becker...

3

maximizers change motivations over time, from a positive reinforcement to a negative reinforcement. If negative reinforcement controls behavior, subjects won't be able to anticipate harmful consequences of current consumption. Stock of consumption capital depends on past consumption of an addictive good and life cycle events. Addictive capital may indirectly depend on how many resources subject has to face an undesired or stressful event. Less resources a subject has, more likely he may become addicted. Resource pool varies according to how many successful experiences of self-control in the past a subject has. Thus, stock of consumption capital may be an interaction between current consumption and current experiences as they both are weighted according to past experiences in memory. Can memory play a role in a theory of rational addiction according to this way of reasoning? Past consumption might be valued as beneficial while subject held in memory a related experience about a stressful event. In this case, any good could become addictive if and only if for the subject is causally related to the resolution of a bad emotional state. Addiction could be valued as beneficial for the current stressful event, so a subject may restore his emotional well-being. Could memory be a mechanism to explain how addiction and life cycle events are related according to the theory of rational addiction? Rate of depreciation on addictive capital, may it vary as a function of consumption of an addictive good? Gary Becker:...a lot of questions... Do you want to stress something? I see... Do you want to enphazise something? Reinforcement and tolerance, memory self-control... the rate of depreciation Let me pick a little bit about some of those issues. ... Memory does play a role in our analysis, we have primarily rate of depreciation of the past, we can interpretet that either as a physiological effect so let's say if I drink rapidly that's physiological effects of drugs or alcohol I ache my body that's one of the determinants of this depreciation but in another we've been memory of the experience, how rapidly do I forget the consequences or the effects of drug use in the past or wherever may be, memory plays a role, we... in a simple model it just plays a simple role combining this physiological effect as jointly determining this depretiation rate... in a more complicated role we've a model of addition, we've two differents types of capital each with different memories of depreciation rates, so, yes it plays a role it could be expanded but we do give it a role I think it is important. Some people wonder, you wonder, you're right, when use addictive goods rather than non addictive goods to cope with stressful events, I am not sure that is only addictive goods that are used to cope with stressful events but we do observe is that addictive goods are hard a way we use so that people become unemployed or a marriage breaks off or they have another reverse we know that they began to drink more heavily, they began to smoke, they may back to using drugs, we do know that , but they also do use another things, they may get exercises some of more, they may go out a little bit more, so it is not only addictive types of goods, there is nothing... and I think it is fair to say there is nothing in the theory itself that implies that addictive goods would be more likely to be used during stressful periods. We used that because it seems to be an observation in what we show... and if we have an observation that is easy to explain why addictive people are not happy people, they may be unhappy people. One pressure... my model was... we... our model implies A as happy because it is utility maximizer, and what we say is you must distinguish utility maximizing from being happy, if I am becoming an addict because I suffered my wife left me, I am unhappy but this may make me less unhappy that if I didn't engage in drugs, so it is not I think these criticisms are totally wrong, I miss the point, but I don't have a good explanation for why addictive goods are more likely to be... some addictive goods are likely to be chosen when people are unhappy, I think is only some goods but I don't know the reason why, that's true. Question: Under what conditions may a subject change from an addiction to another? Under what conditions may a subject change from a harmful addiction to a beneficial addiction even without changes in his metapreferences? Gary Becker:...so you say: "Addiction could be valued as beneficial for the current stressful event, so a subject may restore his emotional well-being". ...and I think that's true... It is a vehicle, so in our model addition is always beneficial and that's by assumption, if people are maximizing the utility is beneficial in the sense that they are always better off by using the addictive good than they would be without using them and the benefit may be it helps one cope with some life cycle difficulty. And you can see that certain goods... that is true, if drinking or drug use so to speak, takes one's mind temporaly off of one's current state because... use when person suffers a reverse, they don't want to think about this state, they engage in activities that prevent their

Interviewing Gary Becker...

4

thinking, so they may drink, they may take drugs, but there are a lot other things, they may take a tranquilizer, and may not be addictives but they may do that, right? so it is a way of coping with the pain that they have and it is not difficult to see why certain types of drugs, you know, addictive goods, help you cope with the pain but other goods help you to cope with the pain too. Question: The parameter time preference for the present (s), is it separable from reinforcement? Could it be caused by negative reinforcement, that is, by the need of reducing pain caused by abstinence? If this is the case, addicted subjects are myopic not as a result of a greater discount rate on the future but as a result of a lesser control on current consequences. If this is correct, the parameter time preference for the present (s) would be partly dependent of consumption of an addictive good. Discount rates on the future are not just given and fixed but they are qualitatively different at different states of addiction, before a subject become addicted and after he/she become addicted. Positive reinforcement will have an important role at the initial state but not later as negative reinforcement will be predominant. This would account for a habit may be raised into an addiction by exposure to the habit itself without taking into consideration other variables. At the very last, stopping addiction could cost more than to remain addicted. A person's consumption of a good continues to increase over time even though he fully anticipates the future because of negative reinforcement. There is another feature of this question, positive reinforcement and negative reinforcement, I mean, can we distinguish drug addition to cope with stressful events from addition to cope with the pain of the addition, the drug effect? Gary Becker: Well, both pains, according to us, if you start to use a drug in order to cope with some stressful events, you will recognize, up to a point, that this will cause you some pain in the future; that's part of the analysis, you recognize, you may be uncertain, you may be not certain but you expect some chance that you will suffer future pain and so in our view what you are trading off is less pain now as coping with the stress now for more pain in the future, that's the pain for the drug, that's the tolerance or these other consequences, right?, so tolerance is a future pain, coping with the stress is the present benefit of less pain now, as you trade them off. It's clear when we state in that way that people who discount the future more heavily abound less worry in future pain of the drugs and more concern about coping with stress, people who discount the future less will put more weight in the future pain, that's what those people will be less likely to use drugs as a way of coping with stress, so discounts then play a major role, that's what we enphasize in our analysis, comparing those two pains, so to speak. And what we show in our work is that people with higher discount rates will be more likely to become addictive to this handful things like drugs and so on and be less likely to be addictive to good types of things like exercises and so forth. And in a successful work I have done with another person here in Chicago K. Mulligan, we show, we, now, instead of taking discount rates for persons as given, we show that itself may be affected by various characteristics, their education, their income, may be even their drug use, drug use may affect, drinking may affect, one's capacity to discount the future, they may make you want to discount the future more heavily, and therefore some drug use may stimulate further drug use, it may become accumulative in impact, and t hat is an effect we didn't go into in an original work but I engaged in a subsequent work on time discounting. Question: Stressful events and self-control of behavior May we talk about addictive and non-addictive personalities? (B-M, 1988, p. 691). Paraphrasing a text of yours, we are not sympathetic to the assumption that harmful good addicts had radically different motivations from everyone else. Does not this distinction between addictive and non-addictive personalities contradict a theory of rational addiction? Gary Becker: So, I would't say that we are contradicting the theory of rational addiction. The theory allows for people's discount rates to differ, I mean, there is no single discount rate for a rational person to have, what we show is that different personalities, at least in terms of the different discount rates, may be influenced by some other characteristics like their wealth, so we try to endogenize that, to explain the different personalities, so, in a sense, I think it is an extension of the model of rational behaviour rather than a contradiction. That's how I look at it anyway. Question: It would seem as if some people were naturally inclined to become addicted before some stressful events and some others were not. All factors which lead to addiction seem to be related to a loss of control of the addicted subject on future events, anxiety, stress, loss of job, death of loved people, etc. Could it not be a failure of self-control o selfdeception what causes a person to become addicted? Gary Becker: I think there is some truth in that. Some of that is consistent with rational behavior. I become a big drug user and the drug is going to control me even if I am perfectly rational in the following sense, that I cannot ignore the cumulative drug use that I have had because that's one of the facts that has determined my behavior now; so if I have

Interviewing Gary Becker...

5

been on drugs for three years, it will be rational for me to take account of that in deciding what I should be doing now because the theory is based upon the consequences in part of the acumulated past behavior for what's rational for me to do now. So, yes I think that's true, you can call that loss of self-control, I just would say that it's means that what's rational is contingent on the history of what I have experienced up untill now; if I've been a drug user for now, what would be rational for me would be different for persons who have not been using drugs, they will have different things that are equally rational for that... and to me that ,it is self control but it's the experience partially determines what it's rational to do now. So if I invested a lot... in economics, it's rational to me to continue to be an economist even while someone who invested a lot being a sociologist, it's rational for him to continue to be a sociologist, we both have selfcontrol but we do different things, and that's how I would explain this... I don't like to say it is loss of self-control ,I prefer to say that it's means that my past experience can exercise a big influence on what is rational for me to do now and in the future, and I think that's just as really intrinsic to understanding what addicted behavior means. ...could self deception caused a person to become addicted? Well, of course, it could the question is... important. Is selfdeception important, is a conflict among personalities...? Some models of lack of self-control argue there is a conflict among... with inner person, among different personalities, we don't have that conflict in our approach, there may be different personalities but they agree on what's rational, what should be done, they agree, it's not in conflict, so... which's a better way to understand drug behavior... that's controversial but I feel, if we believe that we don't need to have conflict with inner person in order to have a rich understanding, I think, of what we see with drug users or smoking and there are many ways... why we may be uncertain wether we are going to be addicted in the future, we may be suffering stress now, you know, there are many, many other reasons why people may become addicted without having to stress that they are either myopic, self-deceiving or have lost self control. -Question: Reinforcement has very different roles in addictions depending of consumption of an addictive good and addictive capital. At the very beginning positive reinforcement may have a major role; after a while, as a subject becomes addicted negative reinforcement should predominate. Utility functions should not be the same in both cases: anticipating future consequences could not be the same as subject believes he controls his addiction than as subject recognizes that he is under the control of an addictive good. At the beginning a rational subject want to cope with stressful events but as he becomes addicted he may be trying to escape from pain caused by the abstinence syndrome. Gary Becker: Well, what you are saying is, you are assuming that initially people don't fully anticipate the consequences of what they are doing I think that is the statement is assuming... because if they knew what state they will be getting in, they will recognize they are under control of the addictive good, there may be various reasons what it could be true but not inconsistent with rational addiction, we are publishing a paper in the JPE (Journal of Political Economy) during 1994, by two people... they take the model that Murphy and I have, they bring in uncertainty, people don't know for sure whether they are going to become addicted, they don't know it for certain, so let's say if I am a young person I start taking, drinking heavy layers* , drugs, many people start this and don't become addicted...so how do I know which type I am? and they show, in a rational framework, it's not surprising, that this could lead some people to become addicted because if they had known beforehand that they would become addicted they woudn't have started; that's perfectly consistent with a rational framework assuming one recognizes there is some uncertainty about the addiction process, that's how I would interpret this paragraph. Reinforcement may have different roles but to the extent I can anticipate these different roles I waived that already in my initial decision. If I have uncertain information I may not fully anticipate what's going to happen, I may become addicted and therefore I regret that I began drug use, but regretting I began drug use in a world of uncertainty is not inconsistent with a rational choice initially, that's how we look at it and they show that formally in this paper.... it is a natural way to extent the rational addiction framework. Question: Peer pressure and self-deception Strong peer pressure has an especially large effect on habitual behavior. Strong peer pressure can convert moderately habitual behavior into what appears to be a strong habit or even an addiction by pressing on demand. However, peer pressure may have a beneficial effect on stopping an addiction, under what conditions this beneficial effects could actually take place? does it influence primarily on demand or affects stock of consumption capital? Social control and preference formation How may the social system contribute to preference formation in a way that some subjects can find rational to become consumers of addictive goods? Bolivian Indians use to say that the coca leaf is food for the Indians and poison for the Yankees, under what social conditions if any addictive goods may become harmful for individuals? Is there any

Interviewing Gary Becker...

6

possible social control for an addictive good such as that instituted by ancient cultures? Could unstable steady states of drug consumption become stable by putting in place a social institution, such as a smoking club in the past? How may traditions affect consumption of addictive goods? Gary Becker: Peer pressure is important, for young people specially , strong peer pressure can take what might be moderate addition for each person alone, may make the whole group very addicted. And it is an important phenomena, and it's easy to see why it is more important for addictive goods than for others because many of these addictions, one, it starts when people are teenagers, when we know peer pressure is particularly important, and number two,it is social consumption, consuming in groups, and we know peer pressure is very important for that, so... for those both reasons, we will expect peer pressure to be more important for addictive goods, so I think is a natural extension and we have done some work in my " habit" that actually extend the rational addition model to include peer pressure, and I think that's an inportant direction. Question. And,Can peer preasure have a benefial effect on stopping an adiction? Under what conditions does this take place? Gary Becker: Well, If you can get some members of the peer group to begin to stop, this will influence stopping by other members, another, if you have some members who get influenced... what it means is that you don't have to influence everybody to have a large effect because there will be a snowballing effect from some members to other members through the peer pressure influence, that's the way it can be used. For example, with education policy, if you can educate some of the peers to do so, you can influence the others. That's a right thing can be useful. Question: And, what about this related aspect of the cultural control system? Gary Becker: Culture, I don't know if we can separate that from peer influence. People set the story, I don't know if that is true, about Netherlands having legalized marihuana has led to some reductions in actual use, so it is claimed; I don't know if that's true, and the reason they say that the case is that before, when people were consuming marihuana they did it ,it was illegal, so they did it in hiding groups, small groups where each influences each other to use marihuana, but now it's properly legal, so it does not have to be done in secret, people can go to social clubs an so on, but a lot of people are going to be drug users and there are not users having great influence and it appears now that they did have in the past where those people used more segregated groups because of the early galery; that's an argument which remains to be answered and it is true or not I don't know... it could be true, there are .... cultural forces may have an influence. ...social culture, it could be influential... I don't know the answer to that. Question: C.Prediction failures? According to the theory of rational addiction " the inverse of the number of years of life remaining is an approximation to the rate of "time preference" for people who do not discount the future. Then old people are rationally "myopic" because they have few years of life remaining. Other things the same, therefore, older persons are less concerned about the future consequences of current consumption, and hence they are more likely to become addicted." (B-M, 1988, p. 684). The fact is, according to Nathan (1983) that older people reduce consumption of an addictive good. Bad health could not explain completely this fact. Lower consumption is observed with aging even in cases in which there is no indication of changes in health. It seems reasonable to us that uncertainty about how much life time is left prevents a drug abuser takes into account future consequences. It is likely instead that a better memory of consequences of harmful addictions be a better predictor for quitting or reducing an addiction. Why does memory play no role in rational addictive behavior? If memory of past consumption has a role in managing stressful events, older people would be less addictive than younger people without questioning essential assumptions of the theory of rational addiction. Moreover, an effect of memory may be psychologically a more plausible reason to explain addictive behavior. Gary Becker: The question is why does the theory of rational addition imply that all the people will begin to start addictive behavior?, that's the question, we discuss that a little bit in our paper, we mentioned the health force alone, but we didn't mean to imply that the health is the only reason why older people won't start , let me give a couple of examples, older people don't start crime either, and it may say in a theory of rational behavior, older people should become criminal since imprisonement... an so on, won't be such a deterrence, that's obviously wrong, a difficult with the rational theory are other reasons, I think, in both cases there are other reasons, in addition to the health factor say for drug use, in a theory of addition remenber, memory plays a role in the role of the stock, the effect of the stock of past use, if I have not been a drug user for many years my effect of stock of use may be negative because... it's negative memory because I've forgotten all the consequences so we know that in the theory, the lower the stock of the good, addictive capital, the less likely you are to use it, so old people may have no incentive whatover to start an addictive good, because it is true that fewer years remain but it is also true that they have lower stocks, they have no... smoking or

Interviewing Gary Becker...

7

drinking, then the only way I am going to get started it be if there is a sufficiently strongly stressful event that pushes me into the region where I want to start some form of drug or other forms of addictive behavior, but in the absence of that, I won't start, so I will narrow the question, say well what about older people gone trough highly stressful events, situations, do they start drug use? I presume no, but what I don't know is the extent to which they may start drinking or other addictive behavior, I don't know, we know that all people in general do no start these things, but we don't know the extent of some stressful events have a positive effect on some of the old, certain health reasons, if they did, will be something we may expect to see, so whether the old age relationship is a prediction failure, I will deny that at least in terms of present knowledge there are several ways one can explain why the old may not, do not begin drug use... I don't find any more difficulty in understanding than in understanding why the old don't become criminals or don't begin to do a lot of other comsumption. why not?.. Question: D. Public policies on drug addiction Most countries and states have become to realize the importance of implementing a public policy on drug abuse for reducing state budget, for instance in the areas of law enforcement -i.e.: criminal behavior linked to drug abuse- and health -AIDS dissemination- policies; the analysis provided by the theory of rational addiction has important implications for implementing this public policy, in particular, on the demand side. D1. Thinking in the demand side Effect of price on consumption As shown by you and others (Becker, Grossman y Murphy, 1991), legalization would reduce the prices of drugs such as marijuana, heroin, and cocaine and this would, in turn, increase consumption of these harmful addictive substances. In contrast, large increases in prices would reduce long-run demand. All these are aggregated effects over consumer population, should we expect changes in individual behavior according to changes in price? Larger prices could contract demand if they remain high but this mainly affects actual consumers. Incidentally, larger prices may restrain new consumers going to the market; in contrast, lower prices could have two different effects, they may increase consumption of habitual consumers and/or they make more consumers go to the market. Would both effects actually take place? Conventional wisdom would argue that higher prices may contract demand yet other signs, such as increasing crime rates, would show an underlying lack of elasticity. Could this be a sign of a demand not responsive to prices? Some studies have shown that beer consumption is not elastic in Spain, while according to Lederman (1956) liquor consumption is elastic in US but not in Europe. Gary Becker: One issue would be: is drug use responsive to price?. It's been studied not very much for drugs because of lack of data in most countries, there are some studies are going on... but for cigarettes and alcohol in particular, and we have, as I mentioned earlier a study of demand cigarettes, which is a highly addictive good, cigarettes is a more addictive good than alcohol, very few people can smoke a little bit each day, when many people can drink a little bit each week, so cigarettes is a more addictive good than alcohol, and we find a relative sizeable long run response of smoking to the price of cigarettes, our estimate is that in the long run an increase in price by ten per cent lowers smoking by about seven, eight per cent, that's our estimate, that is not a very pretty big effect, although in the first year lower by four per cent, so it's a bigger long effect than a short run effect which is consistent with the model of rational behavior. People have studied demand for alcohol, we are doing studies for alcohol, we also find generally... in the U.S. perhaps it is different in another countries but in the U.S. quite a big response of alcohol to price, specially for teenagers, beer price, teenagers are sentitive to the price, and teenager group is particularly important since that group is that begins drug use, a lot smoking and studies of cigarettes demand and alcohol demand show particularly high response by teenagers to price changes, so my conclusion is yes, they are very sensitive, the price specially for younger people, may be more for addicts than for other people, a little evidence on that, but I don't see any reasons from the studies I know to believe that price is not important variable, in a social policy point of view.. by taxing cigarettes or alcohol you can cut down use and people have been proposing that in various countries, they call sin-taxes to cut down people sins so you raise, put a tax... a cigarette a lot, and we are proposing, the Clinton administration has proposed to do that, so there are relying on the theory that demand is responsive to price, to try to do that, in part. -Question:... And, is it posible to state some differences between countries? Gary Becker: maybe, studies have been seen about European cigarette demand that show a pretty good response to price for cigarette demand although the trouble of most of these studies on alcohol and cigarettes is that they have not used a model of addition, were our model or any other, people just estimate demand function for appearing and addition, whereas rational addition models or some others... so what we've been trying to do, our innovation in our

Interviewing Gary Becker...

8

empirical work is we bring a model of addition as a way to estimate these effects... we are now finishing studying on alcohol, only for the U.S., and we do find very pretty big effects, for both, specially for teenagers, more for cigarettes than for alcohol, but for the heavy drinkers we have found a big effect, that's the outset why some people may think. -Question: Social discrimination A common result of social campaigns against addictions is that many habitual consumers become acquainted with the consequences of harmful additions, but while many discontinue their addictions, others remain pathologically "hooked". The result is that many consumers quit drugs but others even increase their consumption. May be that some campaigns do not reach socially marginated people or may there be other reasons for some consumers to remain addicted to harmful substances despite the information being provided? Could there be any causes other than addiction itself to remain addicted, such as peer pressure on keeping consumption? Gary Becker: Well, I think it is true... may be for various reasons... less educated, that's one interpretation of the evidence, smoking suggest, also, our studies suggest that the level of poor may be more influenced by legalization than the richer, our theory suggest that, they are going to increase their consumption more, so the other group reduce drug use, the other group will have to concentrate, I think, the social campaigns on less educated both because they are more influenced and because maybe the information gets to them with more difficulties, so I think that's what I respond to that, more... in terms of the information spread and so on should be devoted to the poor and the least educated people that will be the implication of this. Question: Law-enforcement and Punishment Drug consumption is responsive to punishment or repression? Why could be punishment so ineffective as not to be used as a tool for controlling addictive behavior? Would this result contradict the theory of rational addiction? Does there be an optimal punishment to deter increases in the severity of crimes by drug abusers? Could "a victim" be punished? Why should we prosecute drug addicts? Since are they unproductive and/or because are they threatening to keep their harmful addictions at any cost? Should we put beneficial addictions under social control in prevision that in the end they may become harmful addictions? We're thinking in cases like those of some religious sects, events like that of the Waco ranch would be avoided. Gary Becker: Punishment is an interesting question I have though about... Question: specially, why should we prosecute drug addicts?... Gary Becker: I am not very sure if we have to prosecute drugs addits. There is a difference between prosecuting drug suppliers and prosecuting the addicts. I don't think we have to prosecute addicts, the reason for doing this is that you will disturb other people for becoming addicts, so it is the incentive effect. The case against doing it is severe, that they are the ones who are in the worst positions and so we are prosecuting people who , presumably, have more difficulties in their lives and so on, let's see, some people say that's unfair, that by making their lives more difficult we may increase the crime of these individuals, and we know drug addicts produce significant fraction of the crime, so I think is a mistake to prosecute the addicts, although the reason for doing it is that it is one way to cut down drug use but, if you want to cut down drug use, I think is more effective to try to prosecute suppliers even that is difficult because when ever you prosecute suppliers you raise the price of... you don't neccesarily reduce their profits on the average, so you increase the profits of those who avoid being caught as one of the... the theory of crime and punishment... so if you start prosecuting suppliers, those who avoid detection may make more profits on their drugs because of those that are detected are punished and suffer, what you do by prosecuting suppliers you raise the street price of drugs and that's the way of putting down drug use... the advantage of that if you want to cut drug use by prosecuting some groups you will cut drug use but you may pay a price for doing so. Question: D2. Thinking in the supply side We have mainly talked about addictions from the demand side, let's take a view on the supply side; given how successful a public campaign against tobacco consumption has been, could we design a policy that prevent suppliers to block a public policy on demand by, for instance, reducing prices?

Interviewing Gary Becker...

9

While Netherlands is re-studying its public policy on drugs, Switzerland have recently retreated from a permissive policy on drug addiction because of the undesired effects of legalization on private drug consumption; addiction increased dramatically and people around Europe flowed into the country to buy drugs banned in other countries. How could we prepare against undesired effects of legalization such as those derived of a general reduction of prices? Should we invest in research programs aimed at developing innocuous addictive substances instead of putting much emphasis in other public policies? Gary Becker: supply has many dimensions and there are many things we can do in the supply side we can take the Norway ,or the Dutch approach, Nederlands approach, maybe legalizing some drugs as they have been relegalized marihuana and I think they are more tolerant, other countries, I realized Swizerland has retreated from a permissive policy... the way I think we could prepare against the undesirable effects of legalization I see it, so many... of legalization, no question will lower price but I believe it will raise consumption, it's posibble that one raises consumption because the peer pressure effect, let's assume, let's take a worst scenario from this point of view that that raises consumption and maybe not negatively... there are two ways of attacking that I think one is put a moderate sin tax on drug use for various reasons, the externalities of drug uses and so on and use the revenue from that tax maybe other things too, concentrate on educating those groups who are more vulnerable in a legalized frame that I think will be the poor and less educated, so concentrating the campaign on these groups will be most valuable to legalization I think that's one of the... so, moderate sin-tax and educational campaign with the revenue we are getting from the sin tax, I think it could be useful in controlling some of the painful effects, one could also do advertising if you wanted, on drugs, as they advertise now, they probably have some effects in the cigarettes area, although not major effects, and they raise the profits of the drug companies because they have not to spend their money in advertising... competitive advertising, so... I think investing in alternatives that are more inocous in addictive substances is important and I will support and we can do some of the revenue get from the sin tax as a means of financing some of that, and I think if we could find something has some of the same effects on consum but didn't have side effects there will be the ideal world, which satisfy the consumer needs for this without putting them in the difficulties that people get into drug use, I think we would see spontaneous big demand for this. Question: From your point of view, how much emphasis should we put in programs aimed at providing drugs free substituting for more harmful drugs, for example free metadona? Gary Becker: In the present environment, where harmful drugs are illegal, I say, well, I think is a case... I don't know enough about substitubility... but I do know some of that and I think is a case... for doing more, the problem is I think, that it doesn't seem to be such a large demand by drug users for this, because we have these programs in the U.S. and I don't think there is a big demand for it so, there must be some reasons that they don't find this so atractive I don't know the answer to that but I think is worthy exploring further to see if that could be a more effective alternative. * G.Becker mentioned an unknown drink.