# Lesson 2: Drugs in History: How Substances Became Licit or Illicit

*Companion-podcast transcript • Sarah & Kiffer*  
*Office Hours episode to listen to after working through the lesson*

---

**Sarah:** Welcome back to Office Hours. I'm Sarah.

**Kiffer:** And I'm Kiffer. Here are two statements about alcohol prohibition in the United States, from the history behind Lesson two, and both are true. While prohibition was in force, the amount Americans drank roughly doubled. And for the whole time it was in force, Americans drank less than they had before it, still about a third less in its later years. Both come from the same set of estimates.

**Sarah:** Question one. Did alcohol prohibition in the United States work?

**Kiffer:** The estimates come from a 1991 paper by the economists Jeffrey Miron and Jeffrey Zwiebel. Nobody recorded illegal drinking, so they worked backward from things that rise and fall with drinking, such as deaths, admissions to mental hospitals and arrests linked to alcohol. If we set drinking before prohibition at one hundred, it fell to about thirty in the first years, then climbed to about sixty to seventy in the later years. Let's use sixty-five, the middle of that range.

**Sarah:** So here's a question for everyone. Going from thirty to sixty-five, what's the percent change? And going from the pre-prohibition level of one hundred to sixty-five, what's the percent change? Take a few seconds, and pause the audio if you'd like more time.

*(Pause)*

**Sarah:** I'll try it. From one hundred down to sixty-five is a drop of thirty-five, so drinking in the later years was thirty-five percent lower than before prohibition. And from thirty up to sixty-five is a rise of thirty-five, so drinking rose thirty-five percent during prohibition.

**Kiffer:** Check that second answer with the formula for percent change. What do you divide the difference by?

**Sarah:** The earlier value. So it's sixty-five minus thirty, which is thirty-five, divided by thirty, times one hundred. That's about one hundred seventeen percent. Okay, so it's a rise of thirty-five points, which is a rise of about one hundred seventeen percent. Drinking a bit more than doubled from its low point.

**Kiffer:** Right. Your first answer worked because its starting value was one hundred. When you start from one hundred, a change in points and a change in percent happen to be the same number. Start from thirty, and they come apart.

**Sarah:** And that's how both of the opening statements get made. Start from the low point and drinking doubled. Start from before prohibition and drinking stayed lower the whole time.

**Kiffer:** So the first thing to ask of any historical figure is where the count starts and why someone chose that point. A supporter of prohibition will start from the years before it. A critic will start from its first years, when drinking was at its lowest.

**Sarah:** Something else bothers me about the method. If drinking was estimated partly from arrests, aren't arrests also a measure of policing?

**Kiffer:** They are, and that's a fair challenge to any estimate built this way. A proxy works only if its link to drinking stays steady. Arrests for drunkenness depend on how hard police are looking, and that could have changed under prohibition. Deaths depend on what people were drinking as well as how much. Using several measures together helps, but every one of them assumes the link held through years when the law, the market and the drinks themselves all changed. How does the Canadian record compare with the American estimates?

**Sarah:** It points the same way, with much less precision. The Canadian Encyclopedia reports that drunkenness and related crime fell under provincial prohibition, while bootlegging rose sharply, illegal drinking places called blind pigs multiplied, and a doctor's prescription became a loophole. By 1930, every province except Prince Edward Island had moved to selling liquor through government stores. Prince Edward Island had adopted prohibition first, in 1901, and kept it until 1948. So government sale under provincial control, which is a form of legal regulation, is what replaced prohibition in Canada.

**Kiffer:** Keep that in mind, because the same choice comes up again with cannabis. So, did prohibition work?

**Sarah:** I'd like to hear how you'd answer that, because I can argue it either way.

**Kiffer:** It depends on the outcome you name. If the outcome is how much people drank, these estimates say prohibition lowered drinking for its whole duration, by less than its first years suggested. If the outcome includes illegal markets, crime and the cost of enforcement, the picture looks worse. Any verdict on whether it worked has to say which outcomes count and how much each one weighs.

**Sarah:** Question two takes us to Canada's first drug laws, so here's a content note. We'll be talking about racist laws, a racist riot in Vancouver in 1907 and a 1922 book that used racist language. If you'd like support, you can call or text the Suicide Crisis Helpline at nine eight eight, and the Hope for Wellness Help Line for Indigenous people is at one eight five five, two four two, three three one zero. Both are open around the clock. Here's the question. If a drug law was born out of racism, what does that tell us about the drug, and about the law?

**Kiffer:** This one is easy to get wrong in two directions. The first mistake runs like this: the Opium Act of 1908 grew out of anti-Asian politics after the Vancouver riot, so opium must have been harmless and the alarm was invented. The second runs like this: smoking opium carried real risks, so the law was justified whatever the motives behind it. Both of them skip a step in the reasoning.

**Sarah:** Start with the first one. Why doesn't a racist origin settle the question of harm?

**Kiffer:** Because the origin of a law and the effects of a drug are separate questions, and you answer them with different evidence. How opium acts on the body is a question for pharmacology, which Lessons three and five take up. Who wrote the law, why, and on whom it fell are questions for history. The criminologist Neil Boyd's point is that the people who passed Canada's first drug laws never stated why opium deserved a criminal ban when alcohol and tobacco didn't. In his words, the moral logic was never articulated.

**Sarah:** And the second mistake?

**Kiffer:** Here's a detail that helps with that one. In the same year, 1908, Parliament passed a separate law on patent medicines. It required their labels to show the opium and morphine they contained, banned cocaine from those remedies, and left the medicines on sale.

**Sarah:** So here's a question for everyone. In 1908, opium in a bottle of patent medicine got a label. Opium for non-medical use, including the smoking opium made in factories in Victoria and Vancouver, got a ban on import, manufacture and sale. If both laws had been written only from evidence about what opium does to the body, what would you have expected them to look like? Take a few seconds.

*(Pause)*

**Sarah:** I'd have expected the same drug to be treated in roughly the same way, or the strictest rules to fall on whichever form did the most harm, with some evidence offered for which form that was.

**Kiffer:** And Mackenzie King's report of 1908 offered no precise evidence of harm. What it stressed was opium smoking among young white men and women and the profits of Chinese merchants. The form of opium that got the criminal ban was the one linked in the public mind with Chinese residents. The forms sold over the counter got a label. You could defend the split as a line between medical and non-medical use, but patent medicines were sold to anyone, with no prescription, so a label was the main check on the opium in them.

**Sarah:** So even if smoking opium carried real risks, the line these two laws drew follows who was associated with each form of opium more closely than it follows any evidence about harm.

**Kiffer:** That's the careful version of the argument. It still leaves open what a fair law about opium would have looked like in 1908. History can tell you to demand the evidence about a drug, and the evidence itself has to come from somewhere else.

**Sarah:** The same care applies to Emily Murphy. I've seen it said that her book got cannabis banned in Canada.

**Kiffer:** It's a popular claim, and the record behind it is thin. The Black Candle came out in 1922 with a chapter that called cannabis a new menace, and cannabis was added to the schedule in 1923. Boyd found no discussion of that addition in the House of Commons at all. The minister said only that there was a new drug in the schedule. So the book and the decision came out of the same climate, and the evidence doesn't show that one caused the other.

**Sarah:** Which is a little uncomfortable, because the causal version makes a tidier story.

**Kiffer:** It does, and that's the reason to be careful with it. Murphy was a police magistrate writing for a national magazine. She spread a rumour of a racial plot while admitting she couldn't confirm it, and she helped build the climate those laws were passed in. All of that is well supported. A direct line from her chapter to the 1923 schedule isn't, and reading critically means holding the claims you agree with to the same standard.

**Sarah:** Question three is about the representation ratio. In the 1921 to 1922 fiscal year, Chinese residents accounted for six hundred thirty-four of the eight hundred fifty-three convictions under the drug laws, while making up less than half of one percent of Canada's population. That gives a ratio of about one hundred sixty-five. I want to know whether that number tells the whole story.

**Kiffer:** Let's start with a question for everyone. Chinese residents were about 0.45 percent of the population in the 1921 census. What's the largest representation ratio this group could possibly have had, even if every drug conviction in the country had been of a Chinese resident? Take a few seconds.

*(Pause)*

**Kiffer:** If every conviction had been of a Chinese resident, their share of convictions would be one, or one hundred percent. One divided by 0.0045 is about two hundred twenty-two. So for a group this small, the ratio can never go above about two hundred twenty-two.

**Sarah:** And the actual value was one hundred sixty-five, which is already three quarters of the way to that ceiling.

**Kiffer:** Right. The representation ratio compares the group's conviction rate with the conviction rate for the whole population, and the whole population includes the group. When a small group accounts for most of the convictions, its own convictions push up the yardstick it's being measured with. So let's compare the group with everyone else directly, using rates per hundred thousand.

**Sarah:** Okay. For Chinese residents, that's six hundred thirty-four convictions divided by about thirty-nine thousand six hundred people, times one hundred thousand. That's about sixteen hundred per hundred thousand.

**Kiffer:** Good. Now do the same calculation for everyone else in the country.

**Sarah:** Everyone else had eight hundred fifty-three minus six hundred thirty-four, which is two hundred nineteen convictions, among about 8.75 million people. That's about 2.5 per hundred thousand. So the rate ratio is sixteen hundred divided by 2.5, which is about six hundred forty.

**Kiffer:** That's right, and the exact census figures give the same answer, about six hundred forty, which is close to four times the representation ratio. Both numbers are correct, and they answer different questions. The representation ratio compares the group with the whole population. The rate ratio compares the group with everyone outside it.

**Sarah:** So when the group is small and most of the convictions are theirs, the representation ratio understates the gap.

**Kiffer:** Yes. Another way to put it is that in that one year there was a drug conviction for about every sixty-two Chinese residents, counting children and adults, and one for about every forty thousand other residents. One caution: these are counts of convictions, so a person convicted twice counts twice.

**Sarah:** Here's where I want to push. A rate ratio of six hundred forty is so large that someone could say it proves Chinese residents used drugs at much higher rates.

**Kiffer:** And that's the trap in this example. A bigger ratio is still a ratio of convictions. Convictions come from where police looked and which offences were easiest to prove. King's own inquiry heard from an opium manufacturer who said he sold more opium to white customers than to Chinese customers. Both ratios are strong evidence about where enforcement fell, and they say very little about who used drugs.

**Sarah:** So a fair report would give both ratios, say what each one compares, and say plainly that they measure enforcement.

**Kiffer:** That's what I'd want to see. It's also the habit to carry into modern data on drug arrests by race in Canadian cities, which Lesson eleven takes up.

**Sarah:** Question four. When is it fair to call a reaction to a drug a moral panic?

**Kiffer:** Let's test the idea on a case from much earlier. Sarah, you've read the proclamation on coffee-houses. Remind everyone what it said and what happened to it.

**Sarah:** On 29 December 1675, King Charles the Second ordered England's coffee-houses to stop serving from 10 January, citing the idle and disaffected persons who gathered in them and the scandalous reports about his government spread there. Then, on 8 January 1676, two days before it was due to take effect, the proclamation was withdrawn after an outcry from the public and from merchants who sold coffee, tea and chocolate.

**Kiffer:** So here's a question for everyone. Using Goode and Ben-Yehuda's five criteria, concern, hostility, consensus, disproportionality and volatility, which ones does the coffee-house case clearly meet, and which one clearly fails? Take a few seconds.

*(Pause)*

**Sarah:** Concern and hostility are right there in the proclamation's language about idle and disaffected persons. Volatility looks like a fit, since the whole order collapsed in about ten days. Consensus is the one that fails, because the public and the merchants pushed back hard enough to reverse it.

**Kiffer:** I agree, and disproportionality is the one we can't judge from this evidence. The king's worry was political, so judging proportion would mean judging how dangerous coffee-house talk was to his government. The comparison with Canada is the revealing part. Coffee-house customers and the merchants who supplied them were numerous and vocal, and the order was withdrawn. The folk devils of the 1920s in Canada were Chinese residents, a group the law was already taxing at the border and would soon exclude almost entirely. These two cases suggest that whether a crackdown lasts can depend a great deal on how much power its targets have.

**Sarah:** That also shows a problem with volatility as a criterion. The Canadian alarm peaked in the early 1920s and then faded, and the laws it produced stayed until the 1960s.

**Kiffer:** That pattern fits the title Panic and Indifference, the 1991 history of Canada's drug laws by Giffen, Endicott and Lambert. The panic is volatile, and the law it leaves behind can last for decades. So I'd treat volatility as a description of the public mood, and look separately at how long the legal response lasted. By that rule, the ten days in the coffee-house case tell us how long the order lasted, and we'd need other evidence to say how fast the alarm faded.

**Sarah:** My bigger worry is disproportionality. Who decides what a realistic appraisal of the situation is? It sounds like the analyst's own opinion presented as a criterion.

**Kiffer:** It is a judgement, so the honest move is to show the work. Calling something a moral panic claims that the reaction was larger than the harm, which means you need evidence on both sides of that comparison. The label can also be misused to wave away a real concern. The Le Dain Commission is a useful model here. It took the risks of cannabis seriously and still concluded that the criminal law was a poor response to possession.

**Sarah:** Which brings us to the part where we disagree. In 1972, the Le Dain majority recommended repealing the offence of simple possession of cannabis, with no legal sale. Marie-Andrée Bertrand recommended legal distribution under government control. I think Bertrand had the better argument.

**Kiffer:** Make the case, and put a few numbers on the table while you're at it.

**Sarah:** Bertrand reported ten thousand six hundred ninety-five convictions for cannabis possession in 1972, and more than five hundred sixty prison sentences for simple possession that year. Dividing ten thousand six hundred ninety-five by five hundred sixty gives about nineteen, so roughly one possession conviction in nineteen ended in prison. She also estimated that a million Canadians had used cannabis in the past year, which works out to a little more than one conviction for every hundred people who had used it.

**Kiffer:** Bertrand gave those figures in the commission's Final Report of 1973, and they show the costs the commission was weighing. What was wrong with the majority's answer to them?

**Sarah:** The majority's answer was to stop convicting people who possess cannabis while leaving the whole supply illegal. Bertrand saw the problem. She wrote that decriminalizing use alone would inevitably expand the illicit market, and that people would go on buying a product with no controls on price or quality. When Canadian provinces ended alcohol prohibition, they opened government liquor stores, which is close to what Bertrand proposed for cannabis.

**Kiffer:** That's a strong point, and I think supply was the weak spot in the majority's position. Where I'd defend the majority is on what was knowable in 1972. Government liquor control came after alcohol had been a mass commodity for centuries. Cannabis had been used in the past year by about one Canadian in twenty-two, if we take Bertrand's own estimate of a million people in a population of about twenty-two million. A legal market meant predicting its effects with very little evidence to go on, and the international conventions, which Lesson eleven covers, placed cannabis under control. Bertrand's forecast about the illegal market was also a prediction made with little data.

**Sarah:** The majority was predicting too. They said outright that repeal would probably lead to some increase in use.

**Kiffer:** They did, and that's the part of their report I find most useful as a model. They wrote down the cost they expected and explained why they judged it smaller than the cost of applying the criminal law to thousands of young people. So I'd say the majority's step was defensible for its time, and you're right that it left the supply question open. In the end Parliament took neither step, and possession stayed a crime until the Cannabis Act came into force on 17 October 2018. Lesson twelve looks at what that change did.

**Sarah:** I'm still more sympathetic to Bertrand than you are. But I think we agree on a rule.

**Kiffer:** I think we do. For any proposal to change a drug law, ask two questions: what happens to the people who possess the drug, and where does the supply come from?

**Sarah:** And before the change is made, write down the costs you expect, the way the Le Dain majority did, so the results can be compared with the prediction later.

**Kiffer:** Let's pull it together with three things to take away.

**Sarah:** First, a historical figure depends on its starting point. Drinking under United States prohibition roughly doubled from its low point and stayed about a third below its earlier level in the later years, and both statements are true. Ask where the count starts and why.

**Kiffer:** Second, the origin of a drug law and the harms of a drug are separate questions. In 1908, Canada treated the same drug differently depending on who was associated with each form, which tells you how the line was drawn. Evidence about the drug itself has to come from pharmacology and health research.

**Sarah:** And third, conviction figures measure enforcement. A representation ratio and a rate ratio can describe the same records very differently, so say what each one compares, and don't read either one as a measure of who used drugs.

**Kiffer:** If you'd like more practice, rework the 1921 to 1922 convictions problem from this episode. Find the representation ratio, the ceiling for a group of that size and the rate ratio compared with everyone else, then write one sentence on what each number shows and what it can't show.

**Sarah:** Next time, it's Lesson three, How Drugs Work, Routes, Pharmacokinetics and Pharmacodynamics, where the course turns from history to how drugs act in the body.

**Kiffer:** Take care, everyone.

**Sarah:** See you in Lesson three.
