I'm not sure that I've ever seen a list of the personal daily cost of antique drug use, though I did manage to stumble across one in a remarkable little pamphlet by Edward C. Jandy called Narcotic Addiction as a Factor in Petty Larcency in Detroit (published by the Narcotic Committee of the Wayne County Medical Society, et al. in November 1937). There's a lot packed into its 23 pages, not the least of which is a pretty sophisticated look at how to examine the costs of drug addiction to the sales economy of that city.
One of the interesting historical bits that emerges from it is a list of the daily cost of the addiction of one of the target study groups--a selection of 43 local addicts with a combined 673 years of addiction (averaging an unholy 15.5 years of addiction/person). There are immediate limitations to this info--for example there is no correlation to the number of years of addiction to the individually-reported daily drug costs--but since this data seems to be fairly rare it does at least give some idea of the strain of usage per person. And what does it mean to spend $5/day on your heroin habit? CPI is useful, but it is better to look at what that figure means in terms of the average salary and costs of basic goods. If you were working back there in a bad spot of the Depression in 1937 the average salary was about $1,700/year, which means that if these addicts were working (and the great majority wasn't) then they would be spending about 1/3% of their annual income per day--or a little more than all of their daily salary--on their everyday habit.
Spending $1,800 a year on drugs on a $1,700 salary leaves not-so-much-room for anything else but crime, and not having any income at all would mean that all of that money would have to be from criminal activities. In another (potentially gross) way of thinking about this expense is by looking at the average salary in 1937 being about 1/30th of what the average American family income is in 2012, so the daily $5 heroin hit would be something like $150 today, which sounds about right. And if you applied that multiplier to some other standard 1937 prices, the numbers are fairly constant from then to now--the big exception being postage stamps (which would be $1.50 for a first class stamp) and gasoline ($6/gallon), both of which would show a decline. Again, that's a very crude approximation, but it does pause. The author then does some tricky and interesting semi-statistical work with the bottom line showing that drug addicts stole a total of about 3% of the total retail sales (of $545 million) in the U.S. That's a big number--in today's economy, which currently stands at about $33 billion in thefts (or 1.5%) that would 3% for just addicts would be an enormous number, twice the national general total which would spike drug losses at $100 billion for theft alone. I'm thinking that these 1937 stats might be a little (or a lot) loose, but it the report still is decently argued and nicely presented though the data might be not-great--and the daily/habit numbers are a fine thing to find.