1 00:00:00,000 --> 00:00:12,240 Welcome to the Cannabis Data Science Meetup Group. 2 00:00:12,240 --> 00:00:13,800 Welcome to October. 3 00:00:13,800 --> 00:00:16,920 So it's going to be a big month. 4 00:00:16,920 --> 00:00:18,880 Croptober, as some call it. 5 00:00:18,880 --> 00:00:23,840 And so we're going to be looking at cannabis data, as always. 6 00:00:23,840 --> 00:00:26,840 And we should have an interesting day today. 7 00:00:26,840 --> 00:00:31,880 So essentially, I thought we could recap some of the work 8 00:00:31,880 --> 00:00:34,200 we've been doing the past few months 9 00:00:34,200 --> 00:00:37,400 and apply it to Massachusetts. 10 00:00:37,400 --> 00:00:42,280 So we've been calculating various economic statistics 11 00:00:42,280 --> 00:00:43,600 here. 12 00:00:43,600 --> 00:00:46,360 And so before we dive into that, let's just go ahead 13 00:00:46,360 --> 00:00:50,040 and let Heather into the group. 14 00:00:50,040 --> 00:00:53,880 And it's good to see you, Heather. 15 00:00:53,880 --> 00:00:56,680 So good to have you. 16 00:00:56,680 --> 00:00:58,040 Howdy. 17 00:00:58,040 --> 00:00:58,560 Good morning. 18 00:00:58,560 --> 00:00:59,880 Howdy, and doko as well. 19 00:00:59,880 --> 00:01:00,640 Nice seeing you again. 20 00:01:00,640 --> 00:01:01,560 And doko. 21 00:01:01,560 --> 00:01:03,800 Not actually seeing you, I guess. 22 00:01:03,800 --> 00:01:05,320 I think both of our cameras haven't 23 00:01:05,320 --> 00:01:07,040 been on the last session. 24 00:01:07,040 --> 00:01:09,720 So it makes two of us. 25 00:01:09,720 --> 00:01:11,000 All right, on mute now. 26 00:01:13,720 --> 00:01:17,120 We can definitely have more back and forth discussions. 27 00:01:17,120 --> 00:01:21,000 So let me know if you just want to talk cannabis data 28 00:01:21,000 --> 00:01:23,160 at any point. 29 00:01:23,160 --> 00:01:25,160 Or if you just want to listen, I can show you 30 00:01:25,160 --> 00:01:28,000 some of the current research I've been doing here 31 00:01:28,000 --> 00:01:30,560 at Canelytics. 32 00:01:30,560 --> 00:01:36,160 So essentially, what I've gathered and gleaned 33 00:01:36,160 --> 00:01:41,080 is people are interested in data because they 34 00:01:41,080 --> 00:01:42,720 need to do analytics. 35 00:01:42,720 --> 00:01:45,520 They need to get good data. 36 00:01:45,520 --> 00:01:49,720 Part of getting good data is curating the data. 37 00:01:49,720 --> 00:01:52,920 So as we've seen, the data can be messy. 38 00:01:52,920 --> 00:01:56,920 You may need to rename data points. 39 00:01:56,920 --> 00:02:00,680 You may need to turn text into numbers. 40 00:02:00,680 --> 00:02:07,120 You may need to handle text in numbers, for example, 41 00:02:07,120 --> 00:02:13,400 less than signs, malformatted entries, double decimal places, 42 00:02:13,400 --> 00:02:14,960 what have you. 43 00:02:14,960 --> 00:02:16,960 So part of that's the data cleaning. 44 00:02:16,960 --> 00:02:20,200 And then we've seen the augmenting of data, 45 00:02:20,200 --> 00:02:24,960 where you supplement your data set with other data points. 46 00:02:24,960 --> 00:02:31,280 So we've seen that, OK, we can grab data points such as GDP, 47 00:02:31,280 --> 00:02:38,520 hours worked, employment from the Federal Reserve, 48 00:02:38,520 --> 00:02:39,840 economic data. 49 00:02:39,840 --> 00:02:43,680 So we can get those data points, add those to our data sets. 50 00:02:43,680 --> 00:02:45,640 We can get the public data. 51 00:02:45,640 --> 00:02:48,080 And we can standardize that. 52 00:02:48,080 --> 00:02:53,840 And so there has to be some value added there. 53 00:02:53,840 --> 00:02:58,200 And so that's, I think, analytics. 54 00:02:58,200 --> 00:03:00,640 What we're sort of scoping out is our niche. 55 00:03:00,640 --> 00:03:05,320 So we've seen, OK, people are doing some analytics. 56 00:03:05,320 --> 00:03:06,600 So we can help there. 57 00:03:06,600 --> 00:03:09,120 We can also help with the data pipelines. 58 00:03:09,120 --> 00:03:12,720 And we can also provide data as a commodity. 59 00:03:12,720 --> 00:03:19,000 So analytics is dipping our toes into offering some of this data 60 00:03:19,000 --> 00:03:24,040 that we are curating to people in the industry. 61 00:03:24,040 --> 00:03:26,160 And likewise, people in the industry 62 00:03:26,160 --> 00:03:29,200 may be interested in offering their data, 63 00:03:29,200 --> 00:03:36,440 albeit perhaps anonymized, to other parties who are interested. 64 00:03:36,440 --> 00:03:42,400 So look for further action in that regard in the coming weeks. 65 00:03:42,400 --> 00:03:44,760 So that's exciting news that I wanted 66 00:03:44,760 --> 00:03:49,760 to let you know about, early access knowledge. 67 00:03:49,760 --> 00:03:54,240 So without further ado, let's get into some of the data 68 00:03:54,240 --> 00:03:57,000 curation that we're doing. 69 00:03:57,000 --> 00:04:10,680 So all right. 70 00:04:10,680 --> 00:04:21,120 So we've seen that, at least for Massachusetts, 71 00:04:21,120 --> 00:04:26,960 you can access their public data, so public statistics, 72 00:04:26,960 --> 00:04:30,480 daily totals, through an API. 73 00:04:30,480 --> 00:04:34,400 And you can get a handful of interesting data points. 74 00:04:34,400 --> 00:04:40,560 So you can get daily sales, daily number of plants, packages, 75 00:04:40,560 --> 00:04:41,080 employees. 76 00:04:46,720 --> 00:04:50,360 And we've been looking at the economic angle. 77 00:04:50,360 --> 00:04:58,920 So just to recap some of the work that we've done. 78 00:04:58,920 --> 00:05:06,800 So just to show you, so just running our analysis 79 00:05:06,800 --> 00:05:10,480 in Python, using a handful of standard Python packages, 80 00:05:10,480 --> 00:05:14,640 stats models, SciPy, pandas, requests, numpy. 81 00:05:18,920 --> 00:05:20,720 These are just some helper functions. 82 00:05:23,880 --> 00:05:27,040 And this is what we've been doing the past few weeks. 83 00:05:27,040 --> 00:05:33,440 So we've basically aggregated the production data. 84 00:05:33,440 --> 00:05:39,320 And we've also added the licensees data and prices. 85 00:05:39,320 --> 00:05:44,880 So we've gone up through this point in previous weeks. 86 00:05:44,880 --> 00:05:48,000 So last week, we worked with this data 87 00:05:48,000 --> 00:05:52,240 to look at the GDP or total sales in Massachusetts. 88 00:05:52,240 --> 00:05:56,920 So remember last week, we noticed 89 00:05:56,920 --> 00:05:58,760 there were some outliers. 90 00:05:58,760 --> 00:06:05,520 So here, I have simply coded anything greater than, 91 00:06:05,520 --> 00:06:13,040 want to say this is, what's this one? 92 00:06:13,040 --> 00:06:16,680 Yes, I've coded anything greater than 10 million sales 93 00:06:16,680 --> 00:06:19,000 in a day as zero. 94 00:06:19,000 --> 00:06:21,600 I was dividing them by 10, but that wasn't quite working. 95 00:06:21,600 --> 00:06:24,840 So there's something going on with those outliers. 96 00:06:24,840 --> 00:06:27,520 So I hate to just throw away the data point. 97 00:06:27,520 --> 00:06:31,560 But that was throwing the analysis off too much. 98 00:06:31,560 --> 00:06:34,680 Likewise, we shouldn't see sales below zero. 99 00:06:34,680 --> 00:06:36,840 So there were a handful of observations 100 00:06:36,840 --> 00:06:38,280 with negative sales. 101 00:06:38,280 --> 00:06:42,680 So we want to figure out what was going on there as well. 102 00:06:42,680 --> 00:06:53,120 Nonetheless, we can get this data and look at the data. 103 00:06:57,480 --> 00:07:03,640 See, we've got data starting from October 15, 2018. 104 00:07:03,640 --> 00:07:10,800 It's daily going up until October 5 of 2021. 105 00:07:10,800 --> 00:07:13,240 So almost a full year's worth of data here. 106 00:07:13,240 --> 00:07:17,040 So real interesting analysis we can do here. 107 00:07:17,040 --> 00:07:22,480 Almost three years of data here. 108 00:07:22,480 --> 00:07:23,640 Already then. 109 00:07:23,640 --> 00:07:26,880 So we're looking at, right? 110 00:07:26,880 --> 00:07:34,200 We were looking at sales. 111 00:07:34,200 --> 00:07:40,160 This is what you would typically see from a time series sales 112 00:07:40,160 --> 00:07:45,040 where you have what's called stochastic trending. 113 00:07:45,040 --> 00:07:53,600 So you see this real sporadic movement, high variability. 114 00:07:53,600 --> 00:07:55,200 There's still a trend. 115 00:07:55,200 --> 00:08:02,640 And as we noted, it looks like there is this period 116 00:08:02,640 --> 00:08:04,760 where reporting stopped. 117 00:08:04,760 --> 00:08:11,560 So that will throw a bias into our analysis. 118 00:08:11,560 --> 00:08:15,480 So we need to take that into consideration. 119 00:08:15,480 --> 00:08:18,160 Unfortunately, I hate to just throw away the data, 120 00:08:18,160 --> 00:08:21,000 because we've got all this valuable data 121 00:08:21,000 --> 00:08:25,280 prior to April of 2020. 122 00:08:25,280 --> 00:08:27,760 So I hate to throw away that data. 123 00:08:31,600 --> 00:08:36,800 So we'll preface that as we move along. 124 00:08:36,800 --> 00:08:40,520 We can go ahead and essentially just aggregate sales 125 00:08:40,520 --> 00:08:43,240 into monthly and quarterly. 126 00:08:43,240 --> 00:08:46,840 So I'll take the monthly average, 127 00:08:46,840 --> 00:08:49,840 because that's how we'll want to look at employees, right? 128 00:08:49,840 --> 00:08:55,760 The average number of employees at a given point in a month 129 00:08:55,760 --> 00:09:01,680 versus sales, where we'll want to aggregate with a sum, 130 00:09:01,680 --> 00:09:05,880 because we want to look at the total sales in a month. 131 00:09:05,880 --> 00:09:13,280 So we can basically create both of these series here. 132 00:09:13,280 --> 00:09:22,720 And we looked at these last week, 133 00:09:22,720 --> 00:09:25,200 but we can look at them again. 134 00:09:25,200 --> 00:09:30,960 So this is monthly sales, what we were calling GDP. 135 00:09:30,960 --> 00:09:35,040 You see the business cycle prefaced with, 136 00:09:35,040 --> 00:09:40,200 we do have missing observations right at this dip. 137 00:09:40,200 --> 00:09:51,520 So that dip may not be as actual as we think it is. 138 00:09:51,520 --> 00:09:56,040 So nonetheless, we can even look at quarterly 139 00:09:56,040 --> 00:10:02,800 if we're interested in a bit smoother of a series here. 140 00:10:02,800 --> 00:10:06,400 All right, so we created those series. 141 00:10:06,400 --> 00:10:09,680 Just going to grab the licensees data, 142 00:10:09,680 --> 00:10:12,800 just have it on hand. 143 00:10:12,800 --> 00:10:18,960 And we also can grab the price data. 144 00:10:18,960 --> 00:10:20,720 So prices. 145 00:10:24,160 --> 00:10:29,920 In economics, you're often essentially calculating 146 00:10:29,920 --> 00:10:35,880 or estimating the price for everything, right? 147 00:10:35,880 --> 00:10:37,640 Everything has a price. 148 00:10:37,640 --> 00:10:43,960 So that's what the competitive equilibrium is, right? 149 00:10:43,960 --> 00:10:48,920 That's when the market supply equals demand, 150 00:10:48,920 --> 00:10:53,200 everything shakes out, prices are determined, 151 00:10:53,200 --> 00:10:58,920 and a certain quantity is bought or not bought 152 00:10:58,920 --> 00:11:01,560 at a certain price. 153 00:11:01,560 --> 00:11:06,520 So we're going to be pricing everything. 154 00:11:06,520 --> 00:11:12,400 So we've got everything in the market, right? 155 00:11:12,400 --> 00:11:18,560 So we've got the prices on the actual good here. 156 00:11:18,560 --> 00:11:22,440 So this is average price per ounce. 157 00:11:22,440 --> 00:11:25,920 Unfortunately, this is probably average price per ounce 158 00:11:25,920 --> 00:11:27,320 of flour. 159 00:11:27,320 --> 00:11:31,640 So this doesn't take into consideration 160 00:11:31,640 --> 00:11:38,240 the myriad of products we have going on here. 161 00:11:38,240 --> 00:11:42,680 So I'm going to go ahead and hedge my analysis in that. 162 00:11:42,680 --> 00:11:47,160 And just once again, just say the work we do here 163 00:11:47,160 --> 00:11:51,680 at the Canvas Data Science Meetup Group, 164 00:11:51,680 --> 00:11:53,320 a lot of proof of concept. 165 00:11:53,320 --> 00:11:56,800 So we're just saying, OK, the data exists. 166 00:11:56,800 --> 00:11:59,120 You can crunch it in this manner. 167 00:11:59,120 --> 00:12:03,520 If you're going to do this formally, 168 00:12:03,520 --> 00:12:06,520 then you're going to want to be a bit more rigorous in your data 169 00:12:06,520 --> 00:12:09,960 cleaning and what have you. 170 00:12:09,960 --> 00:12:13,840 And we may have a new member. 171 00:12:13,840 --> 00:12:20,200 So good to have you, Steve. 172 00:12:20,200 --> 00:12:25,240 We're doing a market analysis of Massachusetts 173 00:12:25,240 --> 00:12:28,080 at the moment using Canvas Data. 174 00:12:28,080 --> 00:12:32,640 So happy to hear about your background 175 00:12:32,640 --> 00:12:34,120 if you want to chime in. 176 00:12:34,120 --> 00:12:36,600 So just feel free to speak up. 177 00:12:36,600 --> 00:12:38,840 Or if not, then I'll just continue 178 00:12:38,840 --> 00:12:41,760 droning on with this presentation. 179 00:12:46,600 --> 00:12:51,040 So essentially, we're saying, OK, market analysis, 180 00:12:51,040 --> 00:12:53,640 we're determining prices in the market. 181 00:12:53,640 --> 00:12:59,040 We're given the price, presumably, of flour. 182 00:12:59,040 --> 00:13:00,400 So that's awesome. 183 00:13:00,400 --> 00:13:05,080 So we have some of our work already done for us. 184 00:13:05,080 --> 00:13:08,720 So we've got the average price per ounce. 185 00:13:08,720 --> 00:13:15,240 Well, we can put this in meaningful quantities here. 186 00:13:15,240 --> 00:13:18,560 So we'll just say, OK, what's the average price per gram? 187 00:13:18,560 --> 00:13:22,400 Given there's approximately 28 grams in an ounce, 188 00:13:22,400 --> 00:13:27,080 then how many are in a 16th of an ounce? 189 00:13:27,080 --> 00:13:30,080 Because that's a common measurement at the retail 190 00:13:30,080 --> 00:13:31,560 level. 191 00:13:31,560 --> 00:13:35,840 An eighth of an ounce, another common retail measurement. 192 00:13:35,840 --> 00:13:38,520 And then a quarter of an ounce, another common retail 193 00:13:38,520 --> 00:13:39,680 measurement. 194 00:13:39,680 --> 00:13:43,480 And so these are just averages, so nothing glamorous. 195 00:13:43,480 --> 00:13:48,480 In reality, you'd expect a quantity of 1.5 grams 196 00:13:48,480 --> 00:13:51,480 of quantity discounts. 197 00:13:51,480 --> 00:13:56,440 So it would be awesome to have more granular price data. 198 00:13:56,440 --> 00:13:59,200 We'll work with what we're given. 199 00:13:59,200 --> 00:14:04,320 I have a saying, when given peanuts, make peanut butter. 200 00:14:04,320 --> 00:14:10,200 So in this case, we'll work with what we're given. 201 00:14:10,200 --> 00:14:16,880 So we can see, OK, what's the average price 202 00:14:16,880 --> 00:14:20,520 per gram? 203 00:14:20,520 --> 00:14:25,040 Let's just hedge every dip here in August of 2020 204 00:14:25,040 --> 00:14:27,480 with there is missing data. 205 00:14:27,480 --> 00:14:32,720 And we just can't consider that month reliable. 206 00:14:32,720 --> 00:14:35,280 So in fact, later on in the analysis, 207 00:14:35,280 --> 00:14:38,600 I'm excluding, when we start running regressions, 208 00:14:38,600 --> 00:14:41,040 I'm excluding August of 2020. 209 00:14:41,040 --> 00:14:43,280 Because it's an outlier. 210 00:14:43,280 --> 00:14:46,200 We're dropping other outliers so that they 211 00:14:46,200 --> 00:14:48,160 don't bias our analysis. 212 00:14:48,160 --> 00:14:53,560 And there's too many things that are wrong with this. 213 00:14:53,560 --> 00:14:57,800 There are too many sources of bias in this month, right? 214 00:14:57,800 --> 00:15:00,600 For starters, we're missing our data. 215 00:15:00,600 --> 00:15:02,280 So that's bad. 216 00:15:09,840 --> 00:15:14,160 And the relative percent difference, what have you. 217 00:15:14,160 --> 00:15:16,240 Enough of that, we've got our price per gram. 218 00:15:19,360 --> 00:15:21,200 And so this is where I'm going to say, OK, 219 00:15:21,200 --> 00:15:30,640 let's start excluding August of 2020. 220 00:15:30,640 --> 00:15:33,040 I'm not sure if this is. 221 00:15:33,040 --> 00:15:35,480 Let's just say, OK, what was the? 222 00:15:41,680 --> 00:15:42,760 Yeah, that'll exclude it. 223 00:15:42,760 --> 00:15:46,680 Yeah, we'll just say, OK, what's the average price 224 00:15:46,680 --> 00:15:51,280 per gram in the past 12 months? 225 00:15:51,280 --> 00:15:57,280 Well, it's about 13.33 or so. 226 00:16:01,200 --> 00:16:06,000 And if you look at just the price in the last year, 227 00:16:06,000 --> 00:16:09,480 price per gram, this guy, and we've got 228 00:16:09,480 --> 00:16:11,520 a visible negative trend here. 229 00:16:11,520 --> 00:16:16,160 Of course, we're just going from 13.5 to 13.1. 230 00:16:16,160 --> 00:16:20,720 But is this something to take into consideration? 231 00:16:20,720 --> 00:16:24,280 Also, if you're more interested in looking at the price per. 232 00:16:24,280 --> 00:16:38,120 You can look at the price per eight, about $46, $47. 233 00:16:38,120 --> 00:16:40,360 I think this is interesting data. 234 00:16:40,360 --> 00:16:44,960 So if you're a retailer, how are you pricing your products? 235 00:16:44,960 --> 00:16:46,800 Are you above average, below average? 236 00:16:50,080 --> 00:16:52,400 It's a benchmark. 237 00:16:52,400 --> 00:16:58,960 Imperfect benchmark, but it's a benchmark nonetheless. 238 00:16:58,960 --> 00:17:03,040 So we can continue our data curation. 239 00:17:03,040 --> 00:17:07,040 So all of this is data curation, because first, we 240 00:17:07,040 --> 00:17:09,240 had to go through the whole rigmarole 241 00:17:09,240 --> 00:17:11,840 of getting the data through the API. 242 00:17:15,520 --> 00:17:19,720 Not that complicated, but you have to do it. 243 00:17:19,720 --> 00:17:24,000 Then you already saw that, OK, we're already 244 00:17:24,000 --> 00:17:26,400 calculating a data point. 245 00:17:26,400 --> 00:17:30,280 We're already having to curate that data point. 246 00:17:30,280 --> 00:17:34,360 There are imperfections in the data. 247 00:17:34,360 --> 00:17:39,080 So I have to buff those out. 248 00:17:39,080 --> 00:17:47,240 And look, we're now creating entirely new series of data. 249 00:17:47,240 --> 00:17:49,240 This is a whole series of data. 250 00:17:49,240 --> 00:17:53,840 We've given daily data, and now we've 251 00:17:53,840 --> 00:17:58,600 created monthly and quarterly series. 252 00:17:58,600 --> 00:18:03,280 And we've indexed it by the date. 253 00:18:03,280 --> 00:18:09,160 So value added, just got licensees data, 254 00:18:09,160 --> 00:18:12,680 haven't really added much value there yet. 255 00:18:12,680 --> 00:18:14,200 We've got the price data. 256 00:18:14,200 --> 00:18:18,800 We've added a, we've created a few new series here. 257 00:18:18,800 --> 00:18:22,520 So that's not nothing. 258 00:18:22,520 --> 00:18:25,920 Moving on, we can now supplement the data. 259 00:18:25,920 --> 00:18:29,200 So first, to supplement it, we have to actually 260 00:18:29,200 --> 00:18:31,280 get the other data series. 261 00:18:31,280 --> 00:18:37,680 So we saw that the Federal Reserve economic data 262 00:18:37,680 --> 00:18:41,560 is a smashingly good source. 263 00:18:41,560 --> 00:18:45,120 Once again, it has its own imperfections, right? 264 00:18:45,120 --> 00:18:49,960 So nothing's perfect, but you can get some interesting data 265 00:18:49,960 --> 00:18:53,160 points here. 266 00:18:53,160 --> 00:19:00,320 The frequency may not be as we'd like it. 267 00:19:00,320 --> 00:19:04,120 And we may not quite have up to real time, right? 268 00:19:04,120 --> 00:19:05,960 So we only have up through August. 269 00:19:05,960 --> 00:19:09,320 But we'll take what we're given. 270 00:19:09,320 --> 00:19:16,240 So we'll supplement our data here, 271 00:19:16,240 --> 00:19:21,120 initialize Federal Reserve client, 272 00:19:21,120 --> 00:19:24,800 specify when we're starting our analysis. 273 00:19:24,800 --> 00:19:29,480 And we can just grab a handful of data sets. 274 00:19:29,480 --> 00:19:31,800 So this is awesome. 275 00:19:31,800 --> 00:19:34,280 So we can grab the labor force. 276 00:19:34,280 --> 00:19:37,800 We can get the total number of employees in Massachusetts. 277 00:19:37,800 --> 00:19:39,640 We can get the population. 278 00:19:39,640 --> 00:19:50,280 And we can go ahead and get the average weekly data. 279 00:19:50,280 --> 00:19:56,960 And get the average weekly wage. 280 00:19:56,960 --> 00:20:02,240 Not sure we'll need it, but we can get it nonetheless. 281 00:20:02,240 --> 00:20:05,680 And then we can get the average weekly hours worked. 282 00:20:05,680 --> 00:20:12,160 So we can get one, two, three, four, five data sets here. 283 00:20:12,160 --> 00:20:13,920 So that's awesome. 284 00:20:13,920 --> 00:20:20,760 So we already have one, two, three data sets. 285 00:20:20,760 --> 00:20:22,920 So we'll add five more. 286 00:20:26,560 --> 00:20:28,560 And so this is what I'm saying. 287 00:20:28,560 --> 00:20:31,960 We're aggregating data and curating it 288 00:20:31,960 --> 00:20:34,360 in order to provide value. 289 00:20:34,360 --> 00:20:40,120 So we're now working with eight data sets. 290 00:20:40,120 --> 00:20:47,160 We have touched each data set to make it a bit more user 291 00:20:47,160 --> 00:20:47,720 friendly. 292 00:20:53,080 --> 00:20:58,800 And we can now start calculating interesting novel statistics. 293 00:20:58,800 --> 00:21:04,880 So we saw that we were just going to define GDP as sales. 294 00:21:04,880 --> 00:21:13,320 So we can compare cannabis GDP to total GDP in Massachusetts. 295 00:21:16,320 --> 00:21:21,760 And this is our analysis from last week. 296 00:21:21,760 --> 00:21:24,040 So I'm not going to get into it too much. 297 00:21:28,040 --> 00:21:31,520 Other than we saw, OK, cannabis as a percent of GDP 298 00:21:31,520 --> 00:21:33,960 is increasing. 299 00:21:33,960 --> 00:21:43,680 And there's a non-negligible amount of GDP per capita. 300 00:21:43,680 --> 00:21:47,920 And so that's a crude measure of the well-being 301 00:21:47,920 --> 00:21:50,000 of the people of Massachusetts. 302 00:21:50,000 --> 00:21:54,600 So if you were going to try to quantify, OK, 303 00:21:54,600 --> 00:21:58,040 from an economic perspective, how much better off 304 00:21:58,040 --> 00:22:03,800 are people in Massachusetts because of permitting 305 00:22:03,800 --> 00:22:05,800 adult-use cannabis? 306 00:22:05,800 --> 00:22:08,600 Well, you can average it as about, OK, 307 00:22:08,600 --> 00:22:13,760 they're about, I think, what did we say? 308 00:22:23,440 --> 00:22:28,600 So now we're saying they're about $120 or so. 309 00:22:28,600 --> 00:22:30,400 I thought we said it was $140 or so. 310 00:22:30,400 --> 00:22:31,840 So we'll have to double check this. 311 00:22:31,840 --> 00:22:35,560 Yeah, they're a little more than $100 a person better off. 312 00:22:35,560 --> 00:22:38,120 And so it may not seem much, but remember, 313 00:22:38,120 --> 00:22:40,200 this is every single person. 314 00:22:40,200 --> 00:22:43,040 This is every child all the way up 315 00:22:43,040 --> 00:22:46,920 to every retired person. 316 00:22:46,920 --> 00:22:49,040 So it adds up. 317 00:22:52,040 --> 00:22:55,240 It adds up. 318 00:22:55,240 --> 00:22:56,680 And it's increasing, right? 319 00:22:56,680 --> 00:23:02,880 So this is just the beginning. 320 00:23:02,880 --> 00:23:08,520 We're going to essentially try to keep adding more data 321 00:23:08,520 --> 00:23:10,800 points and statistics here. 322 00:23:10,800 --> 00:23:20,880 So this is analysis that we did back in March. 323 00:23:20,880 --> 00:23:34,880 And just to give you a refresher on the economics, 324 00:23:34,880 --> 00:23:37,360 we're going back to March because we're 325 00:23:37,360 --> 00:23:42,520 trying to keep using some of these statistics 326 00:23:42,520 --> 00:23:44,280 that we've looked at along the way. 327 00:23:44,280 --> 00:23:53,320 So here we're looking at employees. 328 00:23:53,320 --> 00:23:54,200 Oh, yes. 329 00:23:54,200 --> 00:24:00,000 And we're essentially saying, OK, we're measuring output. 330 00:24:00,000 --> 00:24:01,040 Why? 331 00:24:01,040 --> 00:24:02,960 So that's our cannabis sales. 332 00:24:02,960 --> 00:24:07,600 And we're saying, OK, let's try to match an economic model 333 00:24:07,600 --> 00:24:08,400 here. 334 00:24:08,400 --> 00:24:16,080 So we said, OK, what if output is a function of plants 335 00:24:16,080 --> 00:24:18,720 and labor? 336 00:24:18,720 --> 00:24:22,720 So this is what's called a Cobb-Douglas production function. 337 00:24:22,720 --> 00:24:26,520 So it's just taking capital and labor. 338 00:24:26,520 --> 00:24:34,080 And they're each raised to the power function. 339 00:24:34,080 --> 00:24:40,120 And then A is a production augmenting technology. 340 00:24:40,120 --> 00:24:43,200 And then you've got a random production shock. 341 00:24:43,200 --> 00:24:46,640 So that kind of encapsulates everything else. 342 00:24:46,640 --> 00:24:56,200 So that's policy decisions or climate-related disasters, 343 00:24:56,200 --> 00:25:07,360 like if a hurricane hits or lightning strikes. 344 00:25:07,360 --> 00:25:12,320 So just kind of moving through this quick for a bit more 345 00:25:12,320 --> 00:25:17,400 in-depth, definitely revisit our competitive wage 346 00:25:17,400 --> 00:25:20,080 and competitive interest rate episodes. 347 00:25:20,080 --> 00:25:23,400 So we're basically just saying, OK, sales 348 00:25:23,400 --> 00:25:25,400 is a function of capital and labor. 349 00:25:30,040 --> 00:25:43,560 And this allows us to basically, I'll have to lay out a, 350 00:25:43,560 --> 00:25:49,520 if you look at my website, I've got the capital and labor 351 00:25:49,520 --> 00:25:54,960 rate, how you derive the wage. 352 00:25:54,960 --> 00:25:58,040 So maybe I'll do this in a future presentation. 353 00:25:58,040 --> 00:26:01,920 So essentially, you can show that the wage rate is 354 00:26:01,920 --> 00:26:11,720 a function of output and labor supply, 355 00:26:11,720 --> 00:26:17,080 as well as our labor productivity beta. 356 00:26:17,080 --> 00:26:20,800 So essentially, they're measuring 357 00:26:20,800 --> 00:26:26,160 on how productive labor is in the production function. 358 00:26:26,160 --> 00:26:27,320 This is a little messy. 359 00:26:31,720 --> 00:26:34,560 Let's see if this one was a little cleaner. 360 00:26:39,960 --> 00:26:41,600 Just canvas soil and plants. 361 00:26:41,600 --> 00:26:42,120 Exactly. 362 00:26:42,120 --> 00:26:44,840 And so then we were basically saying, OK, 363 00:26:44,840 --> 00:26:47,560 given our Cobblet-Douglas production function, 364 00:26:47,560 --> 00:26:54,880 you can estimate the parameters with a regression model. 365 00:26:54,880 --> 00:26:58,280 So you just take the log of your equation, 366 00:26:58,280 --> 00:27:01,600 and then you can basically say, OK, log of sales 367 00:27:01,600 --> 00:27:05,000 is equal to a constant. 368 00:27:05,000 --> 00:27:11,520 But you can also estimate the rate of sales 369 00:27:11,520 --> 00:27:15,920 is exogenous to the model, which means it's not 370 00:27:15,920 --> 00:27:18,240 explained by the model. 371 00:27:18,240 --> 00:27:23,480 Then you have our parameters, so a parameter on capital 372 00:27:23,480 --> 00:27:27,960 and our parameter on labor, and then a random shock 373 00:27:27,960 --> 00:27:32,480 that you expect in a regression. 374 00:27:32,480 --> 00:27:39,040 So that's the dirty economics lesson, quick and dirty. 375 00:27:39,040 --> 00:27:43,200 So we'll do that a bit more in depth 376 00:27:43,200 --> 00:27:45,200 and a bit more thoroughly in the future, 377 00:27:45,200 --> 00:27:50,640 because I'm not satisfied with that explanation. 378 00:27:50,640 --> 00:27:57,800 But nonetheless, we'll just keep the steam rolling along for now. 379 00:27:57,800 --> 00:28:03,640 So we can essentially define those variables here. 380 00:28:03,640 --> 00:28:07,200 So we're just saying, OK, why? 381 00:28:07,200 --> 00:28:11,560 Well, that's our monthly sales. 382 00:28:11,560 --> 00:28:18,480 We can say, so we're going to work with monthly data here 383 00:28:18,480 --> 00:28:23,280 just to smooth out some of the stochasticness 384 00:28:23,280 --> 00:28:26,600 of the daily data. 385 00:28:26,600 --> 00:28:31,880 So capital, we're going to proxy capital 386 00:28:31,880 --> 00:28:35,760 as just the number of total plants tracked. 387 00:28:35,760 --> 00:28:37,800 And so this is going to be everything 388 00:28:37,800 --> 00:28:41,040 from vegetative plants to immature plants 389 00:28:41,040 --> 00:28:44,120 to flowering plants to potentially harvested 390 00:28:44,120 --> 00:28:45,800 and destroyed plants. 391 00:28:45,800 --> 00:28:49,600 So is this a perfect measure of capital? 392 00:28:49,600 --> 00:28:52,040 Absolutely not. 393 00:28:52,040 --> 00:28:55,120 We're just using it as a proxy, because you 394 00:28:55,120 --> 00:28:57,040 have to proxy capital with something, 395 00:28:57,040 --> 00:29:00,720 unless you actually have a measure of capital 396 00:29:00,720 --> 00:29:02,760 at all of these firms. 397 00:29:02,760 --> 00:29:08,640 So we'll do the best with what we have. 398 00:29:08,640 --> 00:29:11,280 So once again, hedge would be much better 399 00:29:11,280 --> 00:29:14,720 to have a better measure of capital. 400 00:29:14,720 --> 00:29:19,160 Labor, we're essentially going to define labor 401 00:29:19,160 --> 00:29:22,560 as the total number of hours worked. 402 00:29:22,560 --> 00:29:24,480 And so we'll have to estimate that, 403 00:29:24,480 --> 00:29:27,880 and we'll estimate that hours worked 404 00:29:27,880 --> 00:29:33,760 as the average number of employees times 405 00:29:33,760 --> 00:29:44,120 the average monthly hours worked times 4, 406 00:29:44,120 --> 00:29:52,920 assuming there are four alarm going off. 407 00:29:52,920 --> 00:30:02,200 So hopefully, we again, sorry for that. 408 00:30:02,200 --> 00:30:07,000 It appears there's some sort of notice. 409 00:30:07,000 --> 00:30:08,000 OK. 410 00:30:08,000 --> 00:30:09,000 OK. 411 00:30:09,000 --> 00:30:13,000 Wait one second. 412 00:30:13,000 --> 00:30:14,000 Whenever you get a break. 413 00:30:14,000 --> 00:30:15,000 OK. 414 00:30:15,000 --> 00:30:16,000 OK. 415 00:30:16,000 --> 00:30:21,000 Just again, you can see the numbers. 416 00:30:21,000 --> 00:30:24,400 OK. 417 00:30:24,400 --> 00:30:32,360 So we can define our economic variables, sales, capital, 418 00:30:32,360 --> 00:30:34,360 and labor. 419 00:30:34,360 --> 00:30:42,200 And we were defining labor as the average number 420 00:30:42,200 --> 00:30:47,440 of employees times the average hours worked times 4, 421 00:30:47,440 --> 00:30:54,480 assuming they're about four weeks in a month. 422 00:30:54,480 --> 00:30:59,440 So we can define our economic variables. 423 00:30:59,440 --> 00:31:08,040 And if we just take a quick look at these, 424 00:31:08,040 --> 00:31:14,400 we'll see that, for example, we don't have a good measure 425 00:31:14,400 --> 00:31:19,320 for October of 2018. 426 00:31:19,320 --> 00:31:27,240 And furthermore, because of the Federal Reserve, 427 00:31:27,240 --> 00:31:30,280 they only have data through August. 428 00:31:30,280 --> 00:31:39,680 So we're limited with our measure of labor 429 00:31:39,680 --> 00:31:42,480 through August. 430 00:31:42,480 --> 00:31:45,600 Once again, we may retry this analysis 431 00:31:45,600 --> 00:31:52,280 just using potentially a different measure for labor. 432 00:31:52,280 --> 00:31:55,240 However, I'm trying to essentially calculate 433 00:31:55,240 --> 00:31:57,880 the average hourly wage. 434 00:31:57,880 --> 00:31:59,640 So that's why I'm trying to get everything 435 00:31:59,640 --> 00:32:03,000 into a measure of hours. 436 00:32:03,000 --> 00:32:06,080 So we'll work with what we're given. 437 00:32:06,080 --> 00:32:15,720 And essentially, we'll drop October of 2018. 438 00:32:15,720 --> 00:32:21,880 We'll also drop August of 2020, because it was an outlier 439 00:32:21,880 --> 00:32:24,640 and there's missing data. 440 00:32:24,640 --> 00:32:30,160 And then also drop September and October of 2021, 441 00:32:30,160 --> 00:32:35,040 because we don't quite have recent data for these months 442 00:32:35,040 --> 00:32:37,160 yet. 443 00:32:37,160 --> 00:32:47,520 So we can exclude these missing observations. 444 00:32:47,520 --> 00:32:53,280 And without further ado, we can estimate the economic model. 445 00:32:53,280 --> 00:33:04,000 All right, so this is the regression that we just ran. 446 00:33:04,000 --> 00:33:12,920 We just ran our dependent variable is sales, 447 00:33:12,920 --> 00:33:14,120 the log of sales. 448 00:33:14,120 --> 00:33:25,040 And our dependent variables were the log of capital 449 00:33:25,040 --> 00:33:35,520 and the dependent variable was the log of capital. 450 00:33:35,520 --> 00:33:48,640 The log of capital and the log of labor. 451 00:33:48,640 --> 00:33:49,320 Oh, I see. 452 00:33:52,520 --> 00:33:57,080 And we've estimated our coefficients here. 453 00:34:01,360 --> 00:34:02,080 Excellent. 454 00:34:02,080 --> 00:34:10,160 So this is interesting. 455 00:34:10,160 --> 00:34:14,640 So I normally would expect alpha to be higher than beta, 456 00:34:14,640 --> 00:34:21,200 but this may be contrary to our expectations. 457 00:34:21,200 --> 00:34:27,880 So essentially, we're estimating that alpha is just 458 00:34:27,880 --> 00:34:30,400 the coefficient on the log of capital 459 00:34:30,400 --> 00:34:33,000 and beta is the coefficient on the log of labor. 460 00:34:33,000 --> 00:34:35,440 So we're basically estimating that alpha 461 00:34:35,440 --> 00:34:50,640 is about approximately 0.19 and beta is approximately 0.45. 462 00:34:50,640 --> 00:34:53,360 For constant returns to scale, you 463 00:34:53,360 --> 00:34:57,120 would expect alpha plus beta. 464 00:34:57,120 --> 00:35:05,160 So if you're just looking at the production function here, 465 00:35:05,160 --> 00:35:07,840 so this is just a mathematical concept. 466 00:35:07,840 --> 00:35:11,480 Constant returns to scale, your power coefficients 467 00:35:11,480 --> 00:35:13,960 would be equal to 1. 468 00:35:13,960 --> 00:35:18,080 Decreasing returns to scale, your coefficients 469 00:35:18,080 --> 00:35:18,920 are less than 1. 470 00:35:18,920 --> 00:35:22,640 And then increasing returns to scale, they're greater than 1. 471 00:35:22,640 --> 00:35:25,880 So here, our coefficients are alpha and beta. 472 00:35:25,880 --> 00:35:30,200 So if alpha plus beta equals 1, you'd 473 00:35:30,200 --> 00:35:32,240 have constant returns to scale. 474 00:35:32,240 --> 00:35:36,080 Here, we've got decreasing returns to scale, 475 00:35:36,080 --> 00:35:43,440 which essentially means we'll reach an equilibrium. 476 00:35:48,880 --> 00:35:50,320 All righty then. 477 00:35:50,320 --> 00:35:58,000 So without further ado, let's use these coefficients. 478 00:35:58,000 --> 00:36:00,600 I'm not sure we need these data points, 479 00:36:00,600 --> 00:36:02,880 but let's grab those statistics. 480 00:36:02,880 --> 00:36:06,680 But yes, and I think I had to write. 481 00:36:06,680 --> 00:36:13,360 So beta is actually our last parameter. 482 00:36:13,360 --> 00:36:17,000 And then alpha is our first parameter. 483 00:36:17,000 --> 00:36:18,120 That's right. 484 00:36:18,120 --> 00:36:30,760 So we can actually go ahead and estimate the competitive wage. 485 00:36:30,760 --> 00:36:34,200 So we already looked at the competitive wage in Colorado. 486 00:36:34,200 --> 00:36:36,360 So now we can look at the competitive wage 487 00:36:36,360 --> 00:36:41,080 in Massachusetts and see if they're comparable. 488 00:36:41,080 --> 00:36:42,880 And this could be interesting, right? 489 00:36:42,880 --> 00:36:46,640 So if you're an employee, do you want to go work in Colorado, 490 00:36:46,640 --> 00:36:49,000 or do you want to go work in Massachusetts? 491 00:36:49,000 --> 00:36:51,720 The cost of living may vary, but you 492 00:36:51,720 --> 00:36:54,960 may want to go to where the wage is highest. 493 00:36:54,960 --> 00:36:57,280 Likewise, if you're a lab director, 494 00:36:57,280 --> 00:37:02,120 you may have to set a different wage rate in Massachusetts 495 00:37:02,120 --> 00:37:05,880 than you may in Colorado. 496 00:37:05,880 --> 00:37:18,160 So without further ado, let's estimate what the wage rate is. 497 00:37:18,160 --> 00:37:25,760 So well, this is interesting. 498 00:37:25,760 --> 00:37:28,320 So let's go back to our model here. 499 00:37:28,320 --> 00:37:34,840 So remember that if we're going to declare 500 00:37:34,840 --> 00:37:39,400 that these are statistically different than 0, 501 00:37:39,400 --> 00:37:44,120 our confidence interval has to not include 0. 502 00:37:44,120 --> 00:37:50,880 So as we can see, at the 5% confidence level, 503 00:37:50,880 --> 00:37:58,440 which is pretty confident, well, if you're a frequentist, 504 00:37:58,440 --> 00:38:02,520 so you wouldn't be able to declare that these 505 00:38:02,520 --> 00:38:04,120 are statistically significant. 506 00:38:04,120 --> 00:38:08,840 These are not statistically significantly different than 0 507 00:38:08,840 --> 00:38:12,240 at the 95% confidence level. 508 00:38:12,240 --> 00:38:22,120 The parameter beta appears to be statistically significantly 509 00:38:22,120 --> 00:38:27,640 different than 0 at the 10% confidence level. 510 00:38:27,640 --> 00:38:31,040 And then alpha is we can't conclude 511 00:38:31,040 --> 00:38:35,360 any statistical significance. 512 00:38:35,360 --> 00:38:39,840 Granted, we have a small data set here, 33 observations 513 00:38:39,840 --> 00:38:43,400 after we've excluded the missing observations. 514 00:38:48,440 --> 00:38:51,560 So not the greatest regression in the world. 515 00:38:51,560 --> 00:38:54,120 And so that explains why we can't 516 00:38:54,120 --> 00:38:55,680 be certain about the wage rate. 517 00:38:55,680 --> 00:38:58,960 In fact, we're saying, oh, the wage rate, the minimum 518 00:38:58,960 --> 00:39:01,200 could be less than 0. 519 00:39:01,200 --> 00:39:03,320 Well, we know that's not actually possible. 520 00:39:06,080 --> 00:39:09,320 Well, or is it? 521 00:39:09,320 --> 00:39:13,240 And so this is where you run into interesting situations 522 00:39:13,240 --> 00:39:17,640 in economics where maybe people would be, in some situations, 523 00:39:17,640 --> 00:39:23,040 willing to work for a negative wage rate in that maybe they'll 524 00:39:23,040 --> 00:39:28,600 commute and incur costs greater than their actual wage. 525 00:39:28,600 --> 00:39:29,920 And why would they do that? 526 00:39:29,920 --> 00:39:32,720 Maybe they're trying to seek experience. 527 00:39:32,720 --> 00:39:37,520 So it's possible, but that's sort of going way out on a limb. 528 00:39:37,520 --> 00:39:44,720 So let's just look at just the wage rate here. 529 00:39:44,720 --> 00:39:50,120 And so first thing I noticed is it's 530 00:39:50,120 --> 00:39:52,200 really high at the beginning. 531 00:39:52,200 --> 00:39:56,200 So this is when the market is just starting up. 532 00:39:56,200 --> 00:40:00,240 And from an economics point of view, this makes sense. 533 00:40:00,240 --> 00:40:03,320 The industry just comes online. 534 00:40:03,320 --> 00:40:05,960 There's not a lot of people working in the cannabis 535 00:40:05,960 --> 00:40:06,720 industry. 536 00:40:06,720 --> 00:40:11,280 The marginal product of labor is going to be incredibly high. 537 00:40:11,280 --> 00:40:13,880 So that means the wage rate is going 538 00:40:13,880 --> 00:40:17,600 to be high if it's a competitive industry, 539 00:40:17,600 --> 00:40:21,600 because just adding one employee is 540 00:40:21,600 --> 00:40:23,640 going to add a whole lot of value. 541 00:40:23,640 --> 00:40:26,560 So you're willing to give them a high wage rate. 542 00:40:26,560 --> 00:40:29,320 Adding the second employee is also 543 00:40:29,320 --> 00:40:31,080 going to provide a lot of value. 544 00:40:31,080 --> 00:40:33,320 So you're also willing to pay them a lot. 545 00:40:33,320 --> 00:40:36,600 As more and more employees come into the industry, 546 00:40:36,600 --> 00:40:40,200 each additional employee adds slightly less 547 00:40:40,200 --> 00:40:42,320 marginal productivity. 548 00:40:42,320 --> 00:40:46,560 So you would expect wages to decrease. 549 00:40:46,560 --> 00:40:49,560 And sure enough, that appears what to have been the case. 550 00:40:49,560 --> 00:40:53,440 So as the industry first comes online, 551 00:40:53,440 --> 00:40:55,800 it would make sense for companies 552 00:40:55,800 --> 00:40:59,720 to pay people hand over fist to come and work for them. 553 00:40:59,720 --> 00:41:03,360 So if you're a producer, if you're a retailer, if you're 554 00:41:03,360 --> 00:41:07,840 a lab, you're willing to attract talent. 555 00:41:07,840 --> 00:41:10,960 You want to attract the best salespeople. 556 00:41:10,960 --> 00:41:19,480 You want to attract the best farmers, the best processors, 557 00:41:19,480 --> 00:41:20,800 the best manufacturers. 558 00:41:20,800 --> 00:41:23,600 You want to attract the best chemists, 559 00:41:23,600 --> 00:41:27,000 the best microbiologists, the best managers. 560 00:41:27,000 --> 00:41:30,920 So you may pay them a premium. 561 00:41:30,920 --> 00:41:32,760 You may pay them a higher wage rate 562 00:41:32,760 --> 00:41:36,880 than they're getting paid in their current industry, 563 00:41:36,880 --> 00:41:43,920 because they're going to add a lot of value 564 00:41:43,920 --> 00:41:46,560 to your company in the cannabis space. 565 00:41:46,560 --> 00:41:51,200 So you can pay them a high wage rate. 566 00:41:51,200 --> 00:41:55,760 As time goes on, it looks like things stabilize. 567 00:41:55,760 --> 00:42:02,400 So this is where we're starting to reach an equilibrium. 568 00:42:02,400 --> 00:42:04,000 Have we reached an equilibrium yet? 569 00:42:04,000 --> 00:42:04,600 Hard to say. 570 00:42:08,000 --> 00:42:12,480 So we'll just say, OK, what's the wage 571 00:42:12,480 --> 00:42:20,240 been in the past 12 months? 572 00:42:20,240 --> 00:42:24,760 We're estimating that in the past 12 months, 573 00:42:24,760 --> 00:42:27,920 the competitive wage in the cannabis industry 574 00:42:27,920 --> 00:42:32,680 in Massachusetts is almost $40 an hour on average. 575 00:42:35,440 --> 00:42:38,200 Are people getting paid the competitive wage? 576 00:42:38,200 --> 00:42:40,720 Not necessarily. 577 00:42:40,720 --> 00:42:43,680 So depending on market frictions, 578 00:42:43,680 --> 00:42:48,320 people may not be getting paid their competitive wage. 579 00:42:48,320 --> 00:42:52,360 If people have strong monopoly power, 580 00:42:52,360 --> 00:42:57,720 then they may get paid more than their competitive wage. 581 00:42:57,720 --> 00:43:07,240 So if there is barriers to entry for certain positions, 582 00:43:07,240 --> 00:43:10,600 I'm not sure the regulations in Massachusetts 583 00:43:10,600 --> 00:43:12,080 are going to be in place. 584 00:43:12,080 --> 00:43:14,640 I know in Colorado, you have to get a license 585 00:43:14,640 --> 00:43:16,080 to be in the cannabis industry. 586 00:43:16,080 --> 00:43:18,520 So these are various frictions. 587 00:43:18,520 --> 00:43:22,240 So employees can use these frictions 588 00:43:22,240 --> 00:43:27,360 to ask for a rate above their competitive wage rate. 589 00:43:34,480 --> 00:43:38,040 Conversely, I'll need to think a bit more about this. 590 00:43:38,040 --> 00:43:41,640 If companies have stronger bargaining power, 591 00:43:41,640 --> 00:43:43,840 then they may not be able to get their competitive wage. 592 00:43:43,840 --> 00:43:47,360 I'll need to think more and more on this. 593 00:43:47,360 --> 00:43:52,640 For now, let's just focus on the data here. 594 00:43:52,640 --> 00:43:58,880 So just curious what the max wage. 595 00:44:02,880 --> 00:44:07,840 OK, so we're saying, OK, they could be paid up to $40. 596 00:44:07,840 --> 00:44:12,520 They could be paid up to $51 an hour. 597 00:44:12,520 --> 00:44:14,720 And this is your average employee. 598 00:44:14,720 --> 00:44:19,440 This is your retailer. 599 00:44:19,440 --> 00:44:21,520 There are people working in retail. 600 00:44:21,520 --> 00:44:23,480 But then again, remember, this is 601 00:44:23,480 --> 00:44:26,880 a wildly inaccurate measure because we're saying, 602 00:44:26,880 --> 00:44:30,320 oh, our lower bound is negative 18. 603 00:44:30,320 --> 00:44:32,600 That's not even really possible. 604 00:44:32,600 --> 00:44:39,280 So we've got a large room for errors here. 605 00:44:39,280 --> 00:44:45,240 Just to continue, so remember, let's stay on track here. 606 00:44:45,240 --> 00:44:47,600 So we've gotten the prices. 607 00:44:47,600 --> 00:44:51,040 And then we said, OK, this is about the price per unit. 608 00:44:51,040 --> 00:44:56,720 Well, we just priced labor. 609 00:44:56,720 --> 00:45:02,840 In economics, everything has a price, and in equilibrium, 610 00:45:02,840 --> 00:45:05,280 we determine the price. 611 00:45:05,280 --> 00:45:11,920 So at each point in time, the market reached an equilibrium. 612 00:45:11,920 --> 00:45:15,120 People decided how much labor they were going to supply. 613 00:45:15,120 --> 00:45:18,960 Employers decided how much labor they were going to employ. 614 00:45:18,960 --> 00:45:22,720 The wages were set, and the employees were paid, 615 00:45:22,720 --> 00:45:26,400 and the revenue was gained. 616 00:45:26,400 --> 00:45:32,920 So at each point in time, we've got an equilibrium. 617 00:45:32,920 --> 00:45:35,640 And so the equilibrium competitive wage 618 00:45:35,640 --> 00:45:40,200 will be fluctuating over time. 619 00:45:40,200 --> 00:45:43,520 And that's essentially just the price for people's labor. 620 00:45:43,520 --> 00:45:46,640 Like, how much does an employer have 621 00:45:46,640 --> 00:45:53,800 to pay to get one hour of labor from the average employee? 622 00:45:53,800 --> 00:45:56,520 So that is the price on labor. 623 00:45:56,520 --> 00:45:57,160 So that's a cost. 624 00:46:00,280 --> 00:46:01,640 But once again, we're determining 625 00:46:01,640 --> 00:46:03,120 the prices in equilibrium. 626 00:46:03,120 --> 00:46:04,280 That's the price on labor. 627 00:46:04,280 --> 00:46:06,720 Moving on. 628 00:46:06,720 --> 00:46:08,960 There's actually a price on capital, right? 629 00:46:08,960 --> 00:46:10,560 So the investors, right? 630 00:46:10,560 --> 00:46:13,920 So there were investors that invested 631 00:46:13,920 --> 00:46:19,520 in these cannabis companies, the productions, the processors, 632 00:46:19,520 --> 00:46:20,800 the retailers. 633 00:46:20,800 --> 00:46:25,160 And they're expecting a return on their investments. 634 00:46:25,160 --> 00:46:30,440 So you can think about this as, OK, so all the physical goods 635 00:46:30,440 --> 00:46:38,920 that were invested in, how much return do those yield? 636 00:46:38,920 --> 00:46:44,280 Well, we can estimate what the average competitive rate 637 00:46:44,280 --> 00:46:45,480 of return is. 638 00:46:45,480 --> 00:46:50,000 Once again, the market may not be competitive, right? 639 00:46:50,000 --> 00:46:53,840 People may have patents on certain technologies. 640 00:46:57,360 --> 00:47:03,040 It may be hard to move equipment from point A to point B. 641 00:47:03,040 --> 00:47:08,600 There's a lot of factors that make the industry 642 00:47:08,600 --> 00:47:09,560 uncompetitive. 643 00:47:09,560 --> 00:47:13,800 And so people may have to pay a higher competitive rate 644 00:47:13,800 --> 00:47:17,040 of return risk, right? 645 00:47:17,040 --> 00:47:21,240 There's a non-negligible chance that your capital may get 646 00:47:21,240 --> 00:47:21,760 seized. 647 00:47:24,720 --> 00:47:29,360 Or there may be some serious losses in capital, right? 648 00:47:29,360 --> 00:47:33,040 So if you invest a lot in this specific parcel of land, 649 00:47:33,040 --> 00:47:35,920 and then all of a sudden the zoning changes, 650 00:47:35,920 --> 00:47:40,720 well, you may incur a loss on some of your investments. 651 00:47:40,720 --> 00:47:44,400 So there's a risk premium. 652 00:47:44,400 --> 00:47:47,800 So these may not be factored into the competitive rate 653 00:47:47,800 --> 00:47:49,080 of return. 654 00:47:49,080 --> 00:47:52,520 Nonetheless, we always say A measure 655 00:47:52,520 --> 00:47:57,400 is better than no measure, even if it's wildly inaccurate. 656 00:47:57,400 --> 00:47:59,200 Well, that may not necessarily be true. 657 00:47:59,200 --> 00:48:02,920 So just hedge that this could be wildly inaccurate. 658 00:48:02,920 --> 00:48:05,280 Once again, it's a proof of concept. 659 00:48:05,280 --> 00:48:07,120 So if you're going to do this on your own, 660 00:48:07,120 --> 00:48:12,200 you can repeat this in a more rigorous manner. 661 00:48:12,200 --> 00:48:21,360 So without further ado, let's see if this looks rational. 662 00:48:21,360 --> 00:48:48,240 So it doesn't quite look like a rational number here. 663 00:48:48,240 --> 00:48:51,240 Yes. 664 00:48:51,240 --> 00:48:54,160 Now I'm struggling with units, right? 665 00:48:54,160 --> 00:49:00,160 Because are they expecting a 110% return? 666 00:49:00,160 --> 00:49:03,440 Are they expecting a 10% return? 667 00:49:03,440 --> 00:49:07,920 Or is this a 1,010% return? 668 00:49:07,920 --> 00:49:13,800 So I'm not happy with this measure here. 669 00:49:16,600 --> 00:49:19,000 Let me think about this for a second. 670 00:49:19,000 --> 00:49:26,160 So this is sales per plant. 671 00:49:31,000 --> 00:49:37,240 So I think this measure must be, if given those units, 672 00:49:37,240 --> 00:49:38,520 dollars per plant. 673 00:49:38,520 --> 00:49:46,520 So this is maybe not the best measure of the rate of return. 674 00:49:46,520 --> 00:49:53,960 Because we're basically saying they're expecting, 675 00:49:53,960 --> 00:50:02,680 they're saying, yes, it's not the best measure. 676 00:50:02,680 --> 00:50:04,120 Because we're basically saying, OK, 677 00:50:04,120 --> 00:50:14,520 on average, you can expect $90 per plant per month 678 00:50:14,520 --> 00:50:16,520 for your investment. 679 00:50:16,520 --> 00:50:20,680 What does a plant cost per month? 680 00:50:20,680 --> 00:50:22,160 It's left out of the picture. 681 00:50:25,880 --> 00:50:31,120 So that's why we were kind of struggling with this. 682 00:50:31,120 --> 00:50:36,120 So that's why we were kind of struggling with the units here. 683 00:50:36,120 --> 00:50:37,920 So the units matter. 684 00:50:37,920 --> 00:50:40,760 So I want to revisit this. 685 00:50:40,760 --> 00:50:44,000 And I may think on this during the week. 686 00:50:44,000 --> 00:50:48,800 And we can maybe discuss this more next week. 687 00:50:48,800 --> 00:50:52,320 But I'm wanting to say that this is sales per plant. 688 00:50:52,320 --> 00:51:02,640 So the rate of return is $90 per plant per month. 689 00:51:02,640 --> 00:51:08,040 That is not a very logical measure of the rate of return. 690 00:51:08,040 --> 00:51:10,640 And I may be wildly incorrect. 691 00:51:10,640 --> 00:51:17,840 So I want to think about this measure a bit more. 692 00:51:17,840 --> 00:51:24,920 One thing, so our measure on the rate of return per plant 693 00:51:24,920 --> 00:51:29,000 is not, well, actually, maybe that's OK. 694 00:51:29,000 --> 00:51:30,160 Maybe that's all right. 695 00:51:30,160 --> 00:51:34,280 So the investor says, OK, we're going 696 00:51:34,280 --> 00:51:41,000 to fund 100 plant room or 5,000 plant room. 697 00:51:41,000 --> 00:51:47,640 Then we want $90 per plant back per month. 698 00:51:47,640 --> 00:51:48,640 That sounds pretty high. 699 00:51:51,600 --> 00:51:54,400 But maybe we'll keep visiting this next week. 700 00:51:54,400 --> 00:51:58,280 And maybe next week we'll try to estimate 701 00:51:58,280 --> 00:52:00,840 the cost of production. 702 00:52:00,840 --> 00:52:03,520 So that way we can get this into percentages. 703 00:52:03,520 --> 00:52:09,160 So what's the cost per plant per month? 704 00:52:09,160 --> 00:52:17,560 So that way we could calculate what the, and what, well, 705 00:52:17,560 --> 00:52:23,080 what we could maybe do is look at the sales per plant 706 00:52:23,080 --> 00:52:23,600 per month. 707 00:52:27,320 --> 00:52:30,920 OK, OK, this is how we can do this here. 708 00:52:30,920 --> 00:52:38,400 So the rate of return, OK, right, 709 00:52:38,400 --> 00:52:43,320 is around $90 per plant. 710 00:52:43,320 --> 00:52:54,760 And each plant is grossing around, 711 00:52:54,760 --> 00:52:56,840 well, what is each plant grossing? 712 00:53:02,400 --> 00:53:06,800 Each tract, well, let's just look 713 00:53:06,800 --> 00:53:13,480 at this in the last 12 months. 714 00:53:13,480 --> 00:53:16,960 So in the last 12 months, each plant 715 00:53:16,960 --> 00:53:23,240 has averaged about $475 per month in revenue. 716 00:53:23,240 --> 00:53:48,800 So I wonder, so rate of return divided by the sales per plant, 717 00:53:48,800 --> 00:54:03,880 the sales per plant. 718 00:54:03,880 --> 00:54:05,280 Didn't do something right here. 719 00:54:05,280 --> 00:54:07,000 Let's see why. 720 00:54:09,960 --> 00:54:11,960 Oh, I see, I see, I see. 721 00:54:11,960 --> 00:54:15,480 OK, so we're just backing out alpha right there. 722 00:54:15,480 --> 00:54:23,480 OK, yeah, I'm going to need to think a bit more about this. 723 00:54:23,480 --> 00:54:25,000 So let's see. 724 00:54:25,000 --> 00:54:27,120 So I think we're going to need to think 725 00:54:27,120 --> 00:54:28,480 a bit more about this. 726 00:54:28,480 --> 00:54:29,960 So let's see. 727 00:54:29,960 --> 00:54:35,400 Yeah, I'm going to need to think a bit more about this. 728 00:54:35,400 --> 00:54:57,480 So price per plant. 729 00:54:57,480 --> 00:55:01,000 Well, I guess that's basically it. 730 00:55:01,000 --> 00:55:12,760 So if the cultivator has to pay the investor around $90 per plant 731 00:55:12,760 --> 00:55:38,560 there, rate to the difference, basically. 732 00:55:38,560 --> 00:55:42,880 I want to say that's profit per plant. 733 00:55:42,880 --> 00:55:49,480 So this is sales per plant net, the rate of return. 734 00:55:49,480 --> 00:55:55,240 So for each plant, they're going to be grossing so much. 735 00:55:55,240 --> 00:55:57,640 But then they have to pay back their investor 736 00:55:57,640 --> 00:56:00,000 a certain amount per plant. 737 00:56:00,000 --> 00:56:08,160 So this would be, this is not quite profit per plant, 738 00:56:08,160 --> 00:56:11,520 because you're not including labor costs here. 739 00:56:15,200 --> 00:56:23,000 But we're kind of moving in the direction of the profit 740 00:56:23,000 --> 00:56:24,200 margin per plant. 741 00:56:24,200 --> 00:56:25,760 I'm going to want to think about this. 742 00:56:25,760 --> 00:56:28,440 We've gotten a little messy here towards the end. 743 00:56:28,440 --> 00:56:32,600 So I'm going to tidy this up for next week, 744 00:56:32,600 --> 00:56:36,000 because we're trying to do the market analysis here 745 00:56:36,000 --> 00:56:39,080 to determine the prices at equilibrium. 746 00:56:39,080 --> 00:56:44,120 And we've kind of gotten a crude measure of, OK, 747 00:56:44,120 --> 00:56:47,040 what's the price per plant? 748 00:56:51,040 --> 00:56:54,480 And I want to say that's the rate of return. 749 00:56:54,480 --> 00:56:56,640 So I want to say we've priced it. 750 00:56:56,640 --> 00:57:04,160 I want to say we've just now priced the cost of a plant. 751 00:57:04,160 --> 00:57:06,280 And we've priced the cost of labor. 752 00:57:11,640 --> 00:57:16,200 So yes, so next week, we should be able to estimate profit 753 00:57:16,200 --> 00:57:17,240 margins here. 754 00:57:22,000 --> 00:57:24,480 So why don't we try to do that next week? 755 00:57:24,480 --> 00:57:26,640 Next week, we try to estimate what 756 00:57:26,640 --> 00:57:30,600 are the profit margins in the cannabis 757 00:57:30,600 --> 00:57:36,480 industry in Massachusetts and try to just see, OK, 758 00:57:36,480 --> 00:57:40,840 is the average profit per licensee, 759 00:57:40,840 --> 00:57:43,920 or perhaps per producer or retailer, 760 00:57:43,920 --> 00:57:47,360 is that increasing or decreasing? 761 00:57:47,360 --> 00:57:50,840 My conjecture is just from what I've seen, 762 00:57:50,840 --> 00:57:56,800 markets tend to get more concentrated over time, which 763 00:57:56,800 --> 00:58:03,040 gives the firm's monopoly power. 764 00:58:03,040 --> 00:58:11,960 But I don't know. 765 00:58:11,960 --> 00:58:14,560 So I'm not going to make a conjecture yet. 766 00:58:14,560 --> 00:58:20,000 So are prices, no, no, no, profits rising or falling? 767 00:58:23,440 --> 00:58:24,160 I don't know. 768 00:58:24,160 --> 00:58:26,840 It depends. 769 00:58:26,840 --> 00:58:31,080 Is the industry becoming more competitive over time or not? 770 00:58:35,320 --> 00:58:37,120 Because if the industry is becoming 771 00:58:37,120 --> 00:58:39,440 more competitive over time, then we're 772 00:58:39,440 --> 00:58:43,920 going to expect average profits to be heading towards zero. 773 00:58:43,920 --> 00:58:46,960 So if the industry is becoming more competitive, 774 00:58:46,960 --> 00:58:51,120 then we should see average profits heading to zero. 775 00:58:51,120 --> 00:58:53,480 If they're heading in the other direction, 776 00:58:53,480 --> 00:58:57,800 then there may be forces at play making the industry less 777 00:58:57,800 --> 00:58:59,440 competitive. 778 00:58:59,440 --> 00:59:03,560 So I think that's something to look at next week, hedging 779 00:59:03,560 --> 00:59:06,600 that these are all crude measures. 780 00:59:06,600 --> 00:59:09,920 But I want to say that we've now all determined prices. 781 00:59:09,920 --> 00:59:13,240 We definitely have the prices of the products. 782 00:59:13,240 --> 00:59:17,200 We have a crude measure of the price of labor. 783 00:59:17,200 --> 00:59:22,120 And I want to say we now have the price for a plant. 784 00:59:22,120 --> 00:59:24,200 I'm going to think about this during the week, 785 00:59:24,200 --> 00:59:27,200 and we can revisit this next week. 786 00:59:27,200 --> 00:59:29,600 But I think we may now have prices. 787 00:59:29,600 --> 00:59:41,520 So now we can look at profits and make 788 00:59:41,520 --> 00:59:50,600 a statement about market competitiveness 789 00:59:50,600 --> 00:59:53,440 and how it may be changing over time in Massachusetts. 790 00:59:53,440 --> 00:59:55,680 And we could even potentially compare that 791 00:59:55,680 --> 00:59:57,800 to other states like Colorado. 792 00:59:57,800 --> 01:00:01,360 And so this is what our industrial organization 793 01:00:01,360 --> 01:00:02,880 economics is all about. 794 01:00:02,880 --> 01:00:07,200 So we're saying, OK, how do these policies in Colorado, 795 01:00:07,200 --> 01:00:13,520 how do they affect the outcomes and the structure 796 01:00:13,520 --> 01:00:15,160 and the performance? 797 01:00:15,160 --> 01:00:18,160 Likewise, how do these states compare? 798 01:00:18,160 --> 01:00:20,320 So how do decisions made in Massachusetts, 799 01:00:20,320 --> 01:00:23,720 how do those affect the market versus decisions 800 01:00:23,720 --> 01:00:25,920 made in Colorado? 801 01:00:25,920 --> 01:00:30,160 And maybe other states can learn from decisions 802 01:00:30,160 --> 01:00:32,760 that these states have made. 803 01:00:32,760 --> 01:00:34,640 Last but not least, we've also looked 804 01:00:34,640 --> 01:00:36,160 at inflation in the past. 805 01:00:36,160 --> 01:00:38,920 So you could just end the day by just saying, 806 01:00:38,920 --> 01:00:45,200 oh, we can just see, OK, how are prices changing over time? 807 01:00:45,200 --> 01:00:50,880 Looks like we have this wild outlier here. 808 01:00:50,880 --> 01:00:58,040 But I just think it's interesting to see just how 809 01:00:58,040 --> 01:00:59,880 prices are changing over time. 810 01:01:04,640 --> 01:01:08,280 And the price of cannabis in Massachusetts 811 01:01:08,280 --> 01:01:14,440 may have stabilized here. 812 01:01:14,440 --> 01:01:20,840 So it'd be interesting to see what the trend is here. 813 01:01:20,840 --> 01:01:23,400 So we'll start looking at that next week. 814 01:01:23,400 --> 01:01:26,480 We'll start looking at the trend in prices 815 01:01:26,480 --> 01:01:28,400 and try to look at market competitiveness 816 01:01:28,400 --> 01:01:29,840 in Massachusetts. 817 01:01:29,840 --> 01:01:37,520 So today, we've begun our analysis here, 818 01:01:37,520 --> 01:01:39,880 our market analysis of Massachusetts. 819 01:01:39,880 --> 01:01:42,560 So we've gotten the data. 820 01:01:42,560 --> 01:01:46,400 We've analyzed the data. 821 01:01:46,400 --> 01:01:50,000 We've curated the data and supplemented it 822 01:01:50,000 --> 01:01:51,040 with other data sets. 823 01:01:54,240 --> 01:01:56,960 And we've now begun our analysis. 824 01:01:56,960 --> 01:01:58,600 We've calculated statistics. 825 01:01:58,600 --> 01:02:04,640 And we are now using economic models, so economic theory, 826 01:02:04,640 --> 01:02:10,160 to estimate what the competitive wage may be 827 01:02:10,160 --> 01:02:13,440 or what the competitive interest rate may be, 828 01:02:13,440 --> 01:02:15,560 or the competitive rate of return on plants 829 01:02:15,560 --> 01:02:18,200 may be in Massachusetts. 830 01:02:18,200 --> 01:02:20,720 So real thorough analysis. 831 01:02:20,720 --> 01:02:24,800 And you're welcome to use all of this work 832 01:02:24,800 --> 01:02:25,800 and do it on your own. 833 01:02:25,800 --> 01:02:29,760 So you're welcome to do your own economic analysis 834 01:02:29,760 --> 01:02:34,400 of another state or repeat my analysis 835 01:02:34,400 --> 01:02:39,680 and try to do it better without some of these shortcomings 836 01:02:39,680 --> 01:02:41,240 that I've run into. 837 01:02:41,240 --> 01:02:45,640 So there's a lot of work to be done here. 838 01:02:45,640 --> 01:02:47,600 So real exciting work here. 839 01:02:47,600 --> 01:02:53,280 And as always, we're using economics, data, cannabis 840 01:02:53,280 --> 01:02:58,600 know-how, and science to have fun 841 01:02:58,600 --> 01:03:01,400 and try to make some interesting insights about the cannabis 842 01:03:01,400 --> 01:03:03,640 industry. 843 01:03:03,640 --> 01:03:11,760 So thank you for tuning in here. 844 01:03:11,760 --> 01:03:17,360 And then are there any questions, comments, concerns, 845 01:03:17,360 --> 01:03:20,360 or ideas, what have you? 846 01:03:20,360 --> 01:03:21,600 So feel free to chime in. 847 01:03:21,600 --> 01:03:24,280 On that note, you're always welcome to contact me. 848 01:03:24,280 --> 01:03:27,240 So you can send me a message through meetup.com 849 01:03:27,240 --> 01:03:32,040 slash cannabis data, cannabis hyphen data hyphen science 850 01:03:32,040 --> 01:03:32,520 dot com. 851 01:03:35,120 --> 01:03:39,000 Just do a Google search for Meetup Cannabis Data Science. 852 01:03:39,000 --> 01:03:40,520 You can also check out the archive 853 01:03:40,520 --> 01:03:43,200 for the latest information on cannabis data. 854 01:03:43,200 --> 01:03:46,040 So if you're interested in the cannabis data science 855 01:03:46,040 --> 01:03:48,200 and cannabis data science, you can 856 01:03:48,200 --> 01:03:52,320 also check out the archive on canlyvix.com 857 01:03:52,320 --> 01:03:54,320 forward slash videos. 858 01:03:54,320 --> 01:03:58,480 So you can find the video archive there. 859 01:03:58,480 --> 01:04:03,840 Do a search on GitHub for the cannabis data science. 860 01:04:03,840 --> 01:04:08,480 And you will find the repository where we have all the meetups. 861 01:04:08,480 --> 01:04:12,640 So you can find all the source code and all the data sets. 862 01:04:12,640 --> 01:04:13,160 Awesome. 863 01:04:13,160 --> 01:04:14,040 Awesome. 864 01:04:14,040 --> 01:04:17,640 So it's not clean. 865 01:04:17,640 --> 01:04:20,640 So as you'll see, some of the work we've done, 866 01:04:20,640 --> 01:04:22,080 it's a little ad hoc. 867 01:04:22,080 --> 01:04:24,720 And we're just kind of writing scripts as we go. 868 01:04:24,720 --> 01:04:28,200 But it's there for you to use. 869 01:04:28,200 --> 01:04:32,080 So you can at least grab snippets that we've written. 870 01:04:32,080 --> 01:04:35,720 You can find data sets that we've compiled. 871 01:04:35,720 --> 01:04:40,480 There are notes to where we've found data sources 872 01:04:40,480 --> 01:04:41,480 and what have you. 873 01:04:41,480 --> 01:04:45,000 So by all means, check it out. 874 01:04:45,000 --> 01:04:47,200 Use anything that you find useful. 875 01:04:47,200 --> 01:04:51,000 And definitely reach out if you want 876 01:04:51,000 --> 01:04:55,000 to talk cannabis science or data science in general. 877 01:04:55,000 --> 01:04:55,680 So always happy. 878 01:04:58,400 --> 01:05:01,960 On that note, thank you for attending the Cannabis Data 879 01:05:01,960 --> 01:05:05,240 Science Meetup Group here in Ohio. 880 01:05:05,240 --> 01:05:08,360 So checking out the medicinal market here in Ohio. 881 01:05:08,360 --> 01:05:11,440 And it's exciting. 882 01:05:11,440 --> 01:05:13,760 So visiting a handful of laboratories here. 883 01:05:13,760 --> 01:05:16,720 And this time, we're moving west. 884 01:05:16,720 --> 01:05:19,040 And just going to keep visiting laboratories 885 01:05:19,040 --> 01:05:24,320 and keeping my ear to the ground of what's 886 01:05:24,320 --> 01:05:25,880 happening in the cannabis industry. 887 01:05:25,880 --> 01:05:27,920 And so I'll let everybody know. 888 01:05:27,920 --> 01:05:33,000 And until next week, we're going to be doing a bit more thinking 889 01:05:33,000 --> 01:05:37,120 about the economics of our economy. 890 01:05:37,120 --> 01:05:39,760 Of our market analysis in Massachusetts. 891 01:05:39,760 --> 01:05:44,440 And think about directions for next week. 892 01:05:44,440 --> 01:05:47,760 So until then, feel free to reach out. 893 01:05:47,760 --> 01:05:50,960 And it's been awesome speaking with you all. 894 01:05:50,960 --> 01:05:53,240 So stay productive, everyone.