1 00:00:00,120 --> 00:00:02,680 It's traditionally been understood that financial 2 00:00:02,680 --> 00:00:06,160 markets exist to allocate capital to its most productive 3 00:00:06,160 --> 00:00:08,640 uses. Recently, however, the 4 00:00:08,640 --> 00:00:12,560 definition of a productive use has been expanded to include 5 00:00:12,760 --> 00:00:16,760 betting on whether a reggaeton artist will wear a dress to a 6 00:00:16,760 --> 00:00:19,880 football match. In the past, if you'd wanted to 7 00:00:19,920 --> 00:00:23,960 place a bet on Bad Bunny's wardrobe choices, or perhaps the 8 00:00:23,960 --> 00:00:28,000 precise timing of an Iranian missile strike, you had to find 9 00:00:28,000 --> 00:00:31,560 a man who conducted his business behind a local pub. 10 00:00:31,920 --> 00:00:36,280 Today, the process is a bit smoother to a rather impressive 11 00:00:36,280 --> 00:00:40,040 piece of legal rebranding. What the authorities used to 12 00:00:40,040 --> 00:00:44,400 call gambling is now referred to as trading event contracts. 13 00:00:44,600 --> 00:00:47,480 This means that when you inevitably lose your money on 14 00:00:47,480 --> 00:00:51,000 these platforms, you're no longer a degenerate gambler, 15 00:00:51,200 --> 00:00:54,880 you're a retail liquidity provider contributing to a truth 16 00:00:54,880 --> 00:00:57,200 machine. Which sounds much more 17 00:00:57,200 --> 00:00:59,800 respectable. I wouldn't necessarily put it on 18 00:00:59,800 --> 00:01:03,600 your CV though. Reading up on prediction markets 19 00:01:03,600 --> 00:01:06,520 is somewhat exhausting, unfortunately. 20 00:01:06,600 --> 00:01:10,560 We seem to be determined to financialize every single aspect 21 00:01:10,560 --> 00:01:13,560 of human existence. It used to be that if you wanted 22 00:01:13,560 --> 00:01:16,640 to know if it might rain tomorrow, you'd just look out 23 00:01:16,640 --> 00:01:20,560 the window, but now it seems you have to check the bid ask spread 24 00:01:20,560 --> 00:01:24,800 on a precipitation swap. The advocates for platforms like 25 00:01:24,800 --> 00:01:29,320 Calcium Polymarket argued that they're providing A vital public 26 00:01:29,320 --> 00:01:32,000 good. The theory is that by allowing 27 00:01:32,000 --> 00:01:35,960 people to bed on anything from congressional control to the 28 00:01:35,960 --> 00:01:40,360 next Federal Reserve decision, the information fed back will be 29 00:01:40,360 --> 00:01:43,120 far more accurate than traditional polling. 30 00:01:43,320 --> 00:01:47,680 Which is a lovely idea. But looking closely at how these 31 00:01:47,680 --> 00:01:50,920 markets actually work, it appears that we haven't so much 32 00:01:50,920 --> 00:01:55,560 invented a truth machine as put a glossy user interface on a 33 00:01:55,560 --> 00:02:00,360 1920s betting shop and to make it even better, invited a group 34 00:02:00,360 --> 00:02:04,000 of quantitative algorithms to come in and extract money from 35 00:02:04,000 --> 00:02:06,520 the public. Today we're going to look at how 36 00:02:06,520 --> 00:02:10,720 these markets actually work, the legal absurdities surrounding 37 00:02:10,720 --> 00:02:14,400 them, why the commodities regulator is suddenly interested 38 00:02:14,400 --> 00:02:17,720 in elections, and what whether any of this makes the world a 39 00:02:17,720 --> 00:02:21,680 better place, or if we've just turned the whole world into a 40 00:02:21,680 --> 00:02:24,880 casino. To understand why the US 41 00:02:24,880 --> 00:02:28,000 government is fighting with itself over who gets to regulate 42 00:02:28,000 --> 00:02:31,880 a bet on a football game, we need to go back to the origins 43 00:02:31,880 --> 00:02:35,920 of American financial regulation, and specifically 44 00:02:35,920 --> 00:02:38,800 onions. In the United States, there's a 45 00:02:38,800 --> 00:02:41,880 federal agency called the Commodity Futures Trading 46 00:02:41,880 --> 00:02:45,920 Commission, or the CFTC. They were originally established 47 00:02:45,960 --> 00:02:50,040 to oversee futures contracts on things like wheat and cotton. 48 00:02:50,320 --> 00:02:54,120 The idea was that these markets would allow farmers and factory 49 00:02:54,120 --> 00:02:58,240 owners to hedge their price risks while making sure that 50 00:02:58,240 --> 00:03:01,120 these contracts weren't just illegal gambling. 51 00:03:01,280 --> 00:03:05,120 For a long time, the CFTC applied what it called an 52 00:03:05,120 --> 00:03:09,080 economic purpose test. They only approved contracts 53 00:03:09,080 --> 00:03:12,440 that served some hedging or price discovery function. 54 00:03:12,720 --> 00:03:15,920 But over the years, the exchange exchanges realized that they 55 00:03:15,920 --> 00:03:19,840 could generate significantly more revenue by listing futures 56 00:03:19,840 --> 00:03:22,360 on interest rates and stock indices. 57 00:03:22,720 --> 00:03:26,800 The CFTC went along with this, and the definition of a 58 00:03:26,800 --> 00:03:31,040 commodity future was slowly expanded to include things that 59 00:03:31,040 --> 00:03:33,120 are not really commodities at all. 60 00:03:33,400 --> 00:03:37,160 A few years ago, Bitcoin was even classified as a commodity. 61 00:03:37,360 --> 00:03:40,920 While it's about as far from being a basic raw material, 62 00:03:41,120 --> 00:03:45,000 agricultural product, or physical asset as anything could 63 00:03:45,000 --> 00:03:48,080 be. So almost anything can be 64 00:03:48,080 --> 00:03:51,600 considered a commodity, with one notable exception. 65 00:03:51,880 --> 00:03:56,920 Thanks to the 1958 Onion Futures Act, you can trade futures on 66 00:03:56,920 --> 00:04:01,360 almost anything in America, but it's illegal to trade futures on 67 00:04:01,360 --> 00:04:03,880 onions. This happened because in the 68 00:04:03,880 --> 00:04:09,080 1950s, two traders managed to corner the Chicago onion market, 69 00:04:09,280 --> 00:04:13,080 artificially inflating the price before crashing it entirely. 70 00:04:13,360 --> 00:04:17,200 The chaos led to angry farmers and, naturally, an act of 71 00:04:17,200 --> 00:04:20,000 Congress. So to summarize the current 72 00:04:20,000 --> 00:04:23,560 state of American financial law, you can legally bet on the 73 00:04:23,560 --> 00:04:25,760 outcome of a geopolitical conflict. 74 00:04:25,920 --> 00:04:29,280 You can bet on the future price of an Internet meme token. 75 00:04:29,480 --> 00:04:32,880 And thanks to platforms like Calci, you can now bet on who 76 00:04:32,880 --> 00:04:35,200 will control the United States Congress. 77 00:04:35,400 --> 00:04:39,640 But if you attempt to hedge your exposure to French onion soup, 78 00:04:39,920 --> 00:04:44,000 or really any onion soup, the federal government will step in 79 00:04:44,000 --> 00:04:48,520 to protect the public from you. No soup for you now. 80 00:04:48,520 --> 00:04:52,400 Up until recently, the CFTC has been quite skeptical of what 81 00:04:52,400 --> 00:04:56,880 they call event contracts. They explicitly banned contracts 82 00:04:56,880 --> 00:05:00,800 involving war, terrorism, assassination, and gaming. 83 00:05:01,080 --> 00:05:05,040 They also tried to ban election contracts, arguing quite 84 00:05:05,040 --> 00:05:08,440 sensibly that approving them would effectively turn the 85 00:05:08,440 --> 00:05:11,800 commodities regulator into an election cop. 86 00:05:12,160 --> 00:05:14,800 And this is not an unreasonable concern. 87 00:05:14,960 --> 00:05:18,160 If you allow people to trade contracts on election outcomes, 88 00:05:18,720 --> 00:05:22,640 the CFTC is then responsible for making sure no one is 89 00:05:22,640 --> 00:05:26,640 manipulating those outcomes, which means that they now would 90 00:05:26,640 --> 00:05:30,040 have to worry about whether an election has been rigged or if 91 00:05:30,040 --> 00:05:32,280 foreign governments are getting involved. 92 00:05:32,480 --> 00:05:37,080 The CFTC was set up to make sure that wheat prices are fair. 93 00:05:37,280 --> 00:05:41,160 Election monitoring may be even further outside their mandate 94 00:05:41,160 --> 00:05:44,120 than Bitcoin. One way or another, the 95 00:05:44,120 --> 00:05:47,480 prediction market platforms decided to push back. 96 00:05:47,760 --> 00:05:51,520 They took the regulator to court, arguing that predicting 97 00:05:51,520 --> 00:05:56,080 an election is not gaming, and somewhat surprisingly, a federal 98 00:05:56,080 --> 00:05:59,600 judge agreed with them. Having established the right to 99 00:05:59,600 --> 00:06:03,360 offer election contracts, the prediction market platforms 100 00:06:03,360 --> 00:06:06,080 naturally decided to push their luck. 101 00:06:06,400 --> 00:06:08,800 Their logic is fairly straightforward. 102 00:06:08,920 --> 00:06:12,560 If a presidential election is just an event, then a football 103 00:06:12,720 --> 00:06:15,960 game is surely also just an event too. 104 00:06:16,320 --> 00:06:20,320 So earlier this year Kalci began self certifying sports 105 00:06:20,320 --> 00:06:23,160 contracts. They started offering bets on 106 00:06:23,160 --> 00:06:26,000 the Super Bowl, the NBA and the Masters. 107 00:06:26,240 --> 00:06:30,400 I say bets, but Kalci would obviously prefer if I said event 108 00:06:30,400 --> 00:06:33,280 contracts. The distinction is important to 109 00:06:33,280 --> 00:06:38,080 them, if not to anyone else. This came as something of a 110 00:06:38,080 --> 00:06:41,120 surprise to the individual American states. 111 00:06:41,280 --> 00:06:45,040 Since the Supreme Court struck down the federal ban on sports 112 00:06:45,040 --> 00:06:49,400 betting in 2018, nearly 40 states have spent a great deal 113 00:06:49,400 --> 00:06:53,760 of time and money setting up highly regulated, heavily taxed 114 00:06:53,760 --> 00:06:57,520 sports betting regimes. They have set up state gambling 115 00:06:57,520 --> 00:07:01,400 licenses, compliance departments, tax collection, the 116 00:07:01,400 --> 00:07:04,520 whole thing. No sooner was that done when 117 00:07:04,520 --> 00:07:07,560 these tech companies turned up offering what appears to be 118 00:07:07,560 --> 00:07:11,600 functionally identical bets, but they're claiming that they're 119 00:07:11,600 --> 00:07:16,440 completely exempt from state gambling laws and state taxes 120 00:07:16,600 --> 00:07:19,960 because their bets are technically federal commodity 121 00:07:19,960 --> 00:07:22,880 swaps. If you are a state gaming 122 00:07:22,880 --> 00:07:26,120 regulator who has spent the last six years building out a 123 00:07:26,120 --> 00:07:29,840 licensing framework, you might find this argument somewhat 124 00:07:29,840 --> 00:07:33,040 frustrating. A number of states have 125 00:07:33,040 --> 00:07:36,200 responded by sending cease and desist letters. 126 00:07:36,520 --> 00:07:40,400 Arizona has gone a step further and simply filed criminal 127 00:07:40,400 --> 00:07:44,280 charges against Kalci for operating an illegal gambling 128 00:07:44,280 --> 00:07:48,160 business, which is a bit of a step up from a cease and desist 129 00:07:48,160 --> 00:07:51,440 letter. Ohio has taken an even more 130 00:07:51,440 --> 00:07:55,120 creative approach. A company there is suing Kalci 131 00:07:55,280 --> 00:08:00,240 using the Statute of Anne. This is a British law from 1710, 132 00:08:00,480 --> 00:08:04,840 passed during the reign of Queen Anne, which allows third parties 133 00:08:04,840 --> 00:08:08,440 to sue to recover other people's gambling losses. 134 00:08:09,040 --> 00:08:13,800 The theory is that if Kalci is operating an illegal sports book 135 00:08:13,800 --> 00:08:18,560 in Ohio, any enterprising person can sue them to get the losing 136 00:08:18,560 --> 00:08:22,280 better's money back. It's the sort of 18th century 137 00:08:22,280 --> 00:08:26,160 legal instrument that you would expect to find in a museum and 138 00:08:26,160 --> 00:08:30,880 not in an active federal lawsuit against AY combinator startup. 139 00:08:31,200 --> 00:08:34,919 But this is where we are now. You might expect the federal 140 00:08:34,919 --> 00:08:39,280 commodities regulator to step in at this point and clarify that a 141 00:08:39,280 --> 00:08:43,520 bet on the New York Knicks is not, in fact, a vital financial 142 00:08:43,520 --> 00:08:45,880 derivative. But they have not. 143 00:08:46,120 --> 00:08:50,440 According to the Financial Times, the CFTC has mostly just 144 00:08:50,440 --> 00:08:54,720 been avoiding the question. Well, it's actually a bit worse 145 00:08:54,720 --> 00:08:57,560 than that. Under new administration, the 146 00:08:57,560 --> 00:09:01,240 CFTC and the Department of Justice have gone to federal 147 00:09:01,240 --> 00:09:05,160 court to block the state of Arizona from enforcing its 148 00:09:05,160 --> 00:09:09,440 gambling laws against Kalci. So the federal government now 149 00:09:09,440 --> 00:09:13,520 appears to be deploying its legal resources to defend a tech 150 00:09:13,520 --> 00:09:18,400 platform's right to operate what Arizona considers an unlicensed 151 00:09:18,400 --> 00:09:22,240 sports book, overriding state law in the process. 152 00:09:22,440 --> 00:09:26,280 Whatever your views may be on prediction markets, you have to 153 00:09:26,280 --> 00:09:30,240 agree that this is a rather unusual use of the Department of 154 00:09:30,240 --> 00:09:34,480 Justice's time. Now, if you're wondering why the 155 00:09:34,480 --> 00:09:37,480 new administration might be so accommodating to these 156 00:09:37,480 --> 00:09:41,440 prediction platforms, there's one small detail that's probably 157 00:09:41,440 --> 00:09:44,720 worth mentioning. A fellow named Donald Trump 158 00:09:44,720 --> 00:09:48,080 Junior, who seems to be some sort of relative of the sitting 159 00:09:48,080 --> 00:09:52,520 president, is currently serving as a strategic advisor to both 160 00:09:52,520 --> 00:09:56,080 Calci and Polymarket. I looked up this fellow's 161 00:09:56,080 --> 00:09:59,760 background and he appears to have no real work experience in 162 00:09:59,800 --> 00:10:04,800 either strategy or advice. He seems to be a reality TV star 163 00:10:04,880 --> 00:10:08,040 who also worked for his dad's real estate company. 164 00:10:08,360 --> 00:10:12,240 I can't think of why they hired him, but I suppose it's still 165 00:10:12,240 --> 00:10:16,520 worth noting that the President son advises the companies that 166 00:10:16,520 --> 00:10:19,560 the federal government is currently shielding from state 167 00:10:19,560 --> 00:10:23,000 prosecutors. I'm sure that it's it's all a 168 00:10:23,000 --> 00:10:26,400 coincidence. Anyhow, if you ask the 169 00:10:26,400 --> 00:10:30,280 executives at these firms why they deserve special treatment 170 00:10:30,280 --> 00:10:33,920 from regulators, they won't talk about the strategic advice 171 00:10:33,920 --> 00:10:36,640 they've received. They'll instead talk about the 172 00:10:36,640 --> 00:10:40,600 truth machine. The core argument in favor of 173 00:10:40,600 --> 00:10:44,240 prediction markets is that they provide more accurate 174 00:10:44,240 --> 00:10:48,840 information than traditional polling or expert analysis does. 175 00:10:49,160 --> 00:10:52,800 The idea is that when people have to back their opinions with 176 00:10:52,800 --> 00:10:55,920 their own money, they strip away their biases. 177 00:10:56,160 --> 00:11:00,160 They bet on what they think will happen, not what they hope will 178 00:11:00,160 --> 00:11:02,960 happen. The idea is that the market 179 00:11:02,960 --> 00:11:06,720 absorbs all available information and prices and 180 00:11:06,720 --> 00:11:09,560 objective probability of an event occurring. 181 00:11:09,880 --> 00:11:13,040 It's the efficient markets hypothesis applied to 182 00:11:13,080 --> 00:11:17,240 everything, not just stock prices, but elections, wars, 183 00:11:17,240 --> 00:11:20,680 weather, and apparently Bad Bunny's wardrobe. 184 00:11:21,040 --> 00:11:25,280 Now this might sound great in a university lecture hall, but the 185 00:11:25,280 --> 00:11:29,280 problem is what happens when these ideas meet reality? 186 00:11:29,520 --> 00:11:33,440 Because prediction markets are often quite small and thinly 187 00:11:33,440 --> 00:11:37,160 traded, it doesn't take too much money to move the odds. 188 00:11:37,360 --> 00:11:41,320 And if the media is treating those odds as objective truth, 189 00:11:41,520 --> 00:11:45,080 then buying the odds might be a really efficient way of buying 190 00:11:45,080 --> 00:11:49,000 positive press coverage. This has happened already. 191 00:11:49,240 --> 00:11:53,000 During the 2012 U.S. presidential election, a single 192 00:11:53,000 --> 00:11:57,880 trader lost around $7,000,000 systematically buying contracts 193 00:11:57,880 --> 00:12:01,600 on Mitt Romney to artificially inflate his chances on a 194 00:12:01,600 --> 00:12:06,040 platform called Intrade. The goal wasn't to win the bet. 195 00:12:06,240 --> 00:12:10,040 The goal was to make the race look closer than it actually was 196 00:12:10,320 --> 00:12:15,320 and to keep voter enthusiasm high, $7,000,000 is a lot of 197 00:12:15,320 --> 00:12:20,320 money to spend on something that isn't a bad, but as a media 198 00:12:20,320 --> 00:12:24,680 strategy, it's not bad. Cable news covered in trades 199 00:12:24,680 --> 00:12:29,360 odds constantly. More recently, in 2021, a 200 00:12:29,360 --> 00:12:33,600 YouTuber named Brian Rose ran for Mayor of London. 201 00:12:34,040 --> 00:12:37,600 During the campaign, he was accused of gaming the odds on 202 00:12:37,600 --> 00:12:42,200 the betting exchange markets by allegedly having people place 203 00:12:42,200 --> 00:12:46,360 bets on his unlikely victory. He could then point to those 204 00:12:46,360 --> 00:12:50,600 betting markets as evidence that he was a serious candidate with 205 00:12:50,600 --> 00:12:54,240 real momentum, which journalists then reported. 206 00:12:54,680 --> 00:12:59,080 So if a wealthy individual or a political campaign can spend a 207 00:12:59,080 --> 00:13:03,200 few $1,000,000 to move the odds on a thinly traded prediction 208 00:13:03,200 --> 00:13:06,920 market, and then point to those odds as evidence of public 209 00:13:06,920 --> 00:13:10,200 support, you haven't really built a truth machine. 210 00:13:10,400 --> 00:13:14,040 You've built APR tool, but one that comes with a chart. 211 00:13:14,520 --> 00:13:18,800 So if prediction markets are not entirely reliable as truth 212 00:13:18,800 --> 00:13:21,440 machines, what are they actually for? 213 00:13:22,120 --> 00:13:25,960 To understand the current boom, it helps to look at the broader 214 00:13:25,960 --> 00:13:29,200 shift in retail investing over the last few years. 215 00:13:29,560 --> 00:13:33,840 Dimitri Kofinis of the Hidden Forces podcast uses the term 216 00:13:33,840 --> 00:13:37,600 financial nihilism to describe what's been going on. 217 00:13:37,920 --> 00:13:42,000 The idea is the traditional paths to building wealth feel 218 00:13:42,000 --> 00:13:45,000 increasingly out of reach for a lot of young people. 219 00:13:45,240 --> 00:13:48,920 So instead of saving and investing carefully, they try to 220 00:13:48,920 --> 00:13:52,040 get rich quickly by putting money into crypto tokens 221 00:13:52,320 --> 00:13:55,800 featuring pictures of dogs that were pitched to them by edgy 222 00:13:55,800 --> 00:13:59,960 billionaires or buying shares in bankrupt companies. 223 00:14:00,680 --> 00:14:03,440 Prediction markets slot in perfectly here. 224 00:14:03,600 --> 00:14:07,320 If you go back five years, crypto was the exciting product 225 00:14:07,480 --> 00:14:11,480 that everyone was talking about. But crypto is kind of dull 226 00:14:11,480 --> 00:14:15,440 today. Bitcoin is up about 25% / 5 227 00:14:15,440 --> 00:14:19,960 years, which sounds OK until you realize that a money market fund 228 00:14:19,960 --> 00:14:24,560 paying 4% with no at all would have gotten you most of the way 229 00:14:24,560 --> 00:14:27,520 there. Your dad has achieved triple the 230 00:14:27,520 --> 00:14:31,440 return of Bitcoin over the last five years with his index fund, 231 00:14:31,600 --> 00:14:35,120 and he didn't have to check his phone at 3:00 in the morning or 232 00:14:35,120 --> 00:14:38,320 pretend to understand what a layer 2 roll up is. 233 00:14:38,800 --> 00:14:42,360 Michael Saylor is desperately trying to make crypto seem 234 00:14:42,360 --> 00:14:46,240 exciting again by structuring leveraged payouts that could 235 00:14:46,240 --> 00:14:50,920 give you a return of 10% or wipe you out entirely, which might 236 00:14:50,920 --> 00:14:53,800 not be the kind of excitement most people need. 237 00:14:54,200 --> 00:14:57,000 At least with prediction markets, you can watch the 238 00:14:57,000 --> 00:15:00,400 sports you've bet on and have something exciting to talk 239 00:15:00,400 --> 00:15:04,280 about, like the fact that someone just bet $100,000 that 240 00:15:04,280 --> 00:15:08,360 the US government will announce the existence of aliens at some 241 00:15:08,360 --> 00:15:12,320 point this year. The problem with all of this is 242 00:15:12,320 --> 00:15:16,240 that whenever a large pool of enthusiastic retail money shows 243 00:15:16,240 --> 00:15:20,520 up somewhere, the professionals are usually not far behind. 244 00:15:21,000 --> 00:15:24,400 According to the Financial Times, large quantitative 245 00:15:24,400 --> 00:15:29,200 trading firms like Susquehanna and DRW, firms that normally act 246 00:15:29,200 --> 00:15:33,160 as market makers on stock exchanges, are now setting up 247 00:15:33,160 --> 00:15:35,640 dedicated prediction market desks. 248 00:15:36,120 --> 00:15:39,320 They're reportedly paying traders base salaries of 249 00:15:39,320 --> 00:15:44,280 $200,000 a year to build algorithms that systematically 250 00:15:44,280 --> 00:15:47,800 identify mispriced contracts on these platforms. 251 00:15:48,080 --> 00:15:51,400 So on one side of the trade, you have a person betting on the 252 00:15:51,400 --> 00:15:55,240 Super Bowl because it seemed like fun, and on the other side 253 00:15:55,240 --> 00:15:58,840 you have a machine that does this 24 hours a day and never 254 00:15:58,840 --> 00:16:02,520 gets excited about anything. This brings us to what the 255 00:16:02,520 --> 00:16:06,160 gambling industry calls the sharks and fish problem. 256 00:16:06,480 --> 00:16:10,720 In the early 2000s, there was a huge boom in online poker. 257 00:16:10,960 --> 00:16:15,120 Millions of amateurs. The fish logged on to play, but 258 00:16:15,120 --> 00:16:18,800 it didn't take long for the professionals or the sharks to 259 00:16:18,800 --> 00:16:21,720 show up. The professionals didn't play 260 00:16:21,720 --> 00:16:23,880 for fun. They played the odds 261 00:16:23,880 --> 00:16:28,000 methodically, and eventually they deployed bots to do it for 262 00:16:28,000 --> 00:16:31,440 them around the clock. The survival time of new 263 00:16:31,440 --> 00:16:35,440 recreational player on these sites was eventually reduced to 264 00:16:35,640 --> 00:16:39,240 not very long. The amateurs worked out that 265 00:16:39,240 --> 00:16:41,800 they were no longer really playing a game. 266 00:16:42,000 --> 00:16:45,640 They were donating their money to a server farm in New Jersey. 267 00:16:45,880 --> 00:16:49,600 They stopped logging in, the liquidity dried up and the whole 268 00:16:49,680 --> 00:16:53,760 ecosystem collapsed. The sharks had eaten all of the 269 00:16:53,760 --> 00:16:58,400 fish and then starved. Today prediction markets are 270 00:16:58,400 --> 00:17:01,960 full of retail money and the platforms are growing quickly. 271 00:17:02,120 --> 00:17:05,920 But unlike trading a meme stock where the price is just whatever 272 00:17:05,920 --> 00:17:09,560 the next person is willing to pay, an event contract 273 00:17:09,560 --> 00:17:12,760 eventually resolves to either true or false. 274 00:17:12,920 --> 00:17:16,960 There is an actual answer. And if you're a retail trader 275 00:17:16,960 --> 00:17:20,920 betting on a geopolitical event based on a feeling, and the 276 00:17:20,920 --> 00:17:24,160 person on the other side of your trade is a gamma neutral 277 00:17:24,160 --> 00:17:28,640 algorithm being run by a multi $50 billion hedge fund, the odds 278 00:17:28,640 --> 00:17:33,040 are not in your favor. This is not a skill gap that can 279 00:17:33,040 --> 00:17:35,120 be closed by doing more research. 280 00:17:35,320 --> 00:17:39,480 It's a structural disadvantage. When the quants have extracted 281 00:17:39,480 --> 00:17:43,600 enough money from the retail public, the excitement will wear 282 00:17:43,600 --> 00:17:45,840 off. The platforms will likely be 283 00:17:45,840 --> 00:17:49,680 left with sports betting, which the states will eventually 284 00:17:49,680 --> 00:17:54,200 either regulate or shut down, and a few novelty contracts that 285 00:17:54,200 --> 00:17:56,920 exist entirely for marketing purposes. 286 00:17:57,160 --> 00:18:00,360 You'll still be able to bet on whether Bad Bunny wears a dress 287 00:18:00,360 --> 00:18:04,240 to a football match, but the truth machine will be mostly 288 00:18:04,240 --> 00:18:07,800 empty. Now, in fairness, there is one 289 00:18:07,800 --> 00:18:10,880 thing prediction markets do better than the traditional 290 00:18:10,880 --> 00:18:13,920 alternative. If you walk into a sports book 291 00:18:13,920 --> 00:18:18,200 or open a DraftKings account and you start winning consistently, 292 00:18:18,360 --> 00:18:22,800 the sports book will reduce your bet size, restrict which markets 293 00:18:22,800 --> 00:18:26,120 you can access, or simply close your account. 294 00:18:26,360 --> 00:18:30,160 This is standard practice. The house is your counterparty, 295 00:18:30,320 --> 00:18:32,800 and a winning better is bad for business. 296 00:18:32,960 --> 00:18:36,560 Several states have tried to pass laws making this illegal, 297 00:18:36,720 --> 00:18:39,240 which tells you how widespread it is. 298 00:18:39,680 --> 00:18:43,840 Prediction markets don't do this because the platform isn't your 299 00:18:43,840 --> 00:18:46,840 counterparty, it's a peer-to-peer exchange. 300 00:18:47,000 --> 00:18:51,040 It just matches buyers and sellers and takes a small fee. 301 00:18:51,400 --> 00:18:54,320 If you're winning, the platform doesn't care. 302 00:18:54,440 --> 00:18:57,600 Someone on the other side of your trade is losing, and the 303 00:18:57,600 --> 00:19:00,120 platform collects its fee either way. 304 00:19:00,360 --> 00:19:04,480 So in that sense, prediction markets are structurally fairer 305 00:19:04,480 --> 00:19:07,480 than sports books. Of course, the reason you're 306 00:19:07,480 --> 00:19:10,680 winning on a prediction market is most likely that you're a 307 00:19:10,680 --> 00:19:15,040 quantitative algorithm, so this is primarily good news for 308 00:19:15,120 --> 00:19:19,000 quantitative algorithms. The traditional sports books 309 00:19:19,160 --> 00:19:21,560 have noticed all of this, by the way. 310 00:19:21,880 --> 00:19:25,400 DraftKings, FanDuel, and Fanatics have all quietly 311 00:19:25,400 --> 00:19:28,520 launched their own prediction market products while 312 00:19:28,520 --> 00:19:33,800 simultaneously spending $48 million on a super PAC to push 313 00:19:33,800 --> 00:19:37,680 for sports betting legalization in states like Texas and 314 00:19:37,680 --> 00:19:40,080 Georgia. So they're fighting prediction 315 00:19:40,080 --> 00:19:43,360 markets with one hand and copying them with the other, 316 00:19:43,520 --> 00:19:47,360 which is a reasonably common strategy in American business. 317 00:19:47,760 --> 00:19:51,600 If you are a retail trader who's tired of losing money to hedge 318 00:19:51,600 --> 00:19:55,040 fund algorithms, you might be tempted to look for an edge. 319 00:19:55,240 --> 00:19:58,320 In traditional financial markets, acquiring non public 320 00:19:58,320 --> 00:20:02,520 material information and trading on it will generally result in a 321 00:20:02,520 --> 00:20:05,600 conversation with the authorities and eventually a 322 00:20:05,600 --> 00:20:07,920 lengthy stay in a federal facility. 323 00:20:08,120 --> 00:20:12,000 In prediction markets, however, insider trading is often 324 00:20:12,000 --> 00:20:16,240 described by proponents as a feature, not a bug. 325 00:20:16,600 --> 00:20:20,440 That's right, a feature. If we look at recent events, 326 00:20:20,640 --> 00:20:23,840 this feature appears to be working remarkably well. 327 00:20:24,080 --> 00:20:28,320 Last summer, a polymarket user operating under the pseudonym 328 00:20:28,520 --> 00:20:33,840 Rico Suave 666 made a series of highly precise and highly 329 00:20:33,840 --> 00:20:37,880 lucrative bets regarding the exact timing of military strikes 330 00:20:37,880 --> 00:20:41,440 in the Middle East. It later turned out that Rico 331 00:20:41,440 --> 00:20:46,480 Suave 666 was not just a very astute reader of geopolitical 332 00:20:46,480 --> 00:20:49,440 tea leaves. The Israeli government arrested 333 00:20:49,440 --> 00:20:53,720 two men, including an army reservist, for allegedly placing 334 00:20:53,720 --> 00:20:56,840 bets using classified military intelligence. 335 00:20:57,280 --> 00:21:01,480 So to be clear, a soldier would advance knowledge of when bombs 336 00:21:01,480 --> 00:21:05,320 were going to be dropped. Use that information to win 337 00:21:05,320 --> 00:21:09,560 money on what is essentially A crypto gambling website, which 338 00:21:09,560 --> 00:21:13,240 is not really what people have in mind when they talk about the 339 00:21:13,240 --> 00:21:17,240 wisdom of crowds. We saw something similar with 340 00:21:17,240 --> 00:21:21,480 the capture of Nicolas Maduro. Shortly after the United States 341 00:21:21,480 --> 00:21:25,840 announced the operation, someone placed a series of very large 342 00:21:25,840 --> 00:21:29,400 and very confident bets on Polymarket that he would be 343 00:21:29,400 --> 00:21:33,840 removed from office, walking away with a few $100,000. 344 00:21:34,040 --> 00:21:37,560 It's not clear who placed those bets, but they appear to have 345 00:21:37,560 --> 00:21:41,600 had a better understanding of US foreign policy than most of the 346 00:21:41,840 --> 00:21:45,200 US Senate. Now, if you ask the operators 347 00:21:45,200 --> 00:21:49,200 and advocates of these markets about this sort of thing, they 348 00:21:49,200 --> 00:21:53,440 are surprisingly relaxed. They argue that insider trading 349 00:21:53,440 --> 00:21:57,160 is actually a good thing. The logic is that the insider 350 00:21:57,160 --> 00:22:00,880 brings valuable information to the market, which makes the 351 00:22:00,880 --> 00:22:05,040 price more accurate. The market absorbs the leak, the 352 00:22:05,040 --> 00:22:09,200 odds adjust, and society gets a more accurate forecast. 353 00:22:09,680 --> 00:22:13,560 This, they say, is the truth machine working exactly as 354 00:22:13,560 --> 00:22:17,080 intended. It's a rather creative argument. 355 00:22:17,280 --> 00:22:21,600 If a military officer leaks classified operational plans so 356 00:22:21,600 --> 00:22:25,320 that his friend can win a few $100,000 on a crypto betting 357 00:22:25,320 --> 00:22:29,040 site, we should apparently all be grateful for the positive or 358 00:22:29,040 --> 00:22:32,840 externality of slightly more accurate price discovery. 359 00:22:33,080 --> 00:22:36,240 I'm sure that the soldiers involved in those operations 360 00:22:36,400 --> 00:22:40,200 would be reassured to learn that their safety was compromised in 361 00:22:40,200 --> 00:22:42,840 the noble pursuit of market efficiency. 362 00:22:43,160 --> 00:22:47,360 The reason insider trading is banned in the stock market is 363 00:22:47,360 --> 00:22:50,560 fairly simple. If ordinary investors believe 364 00:22:50,560 --> 00:22:54,960 the game is rigged, that only insiders can win, they will stop 365 00:22:54,960 --> 00:22:57,640 investing. And if people stop investing, 366 00:22:57,840 --> 00:23:01,880 companies can't raise capital to build factories, fund research, 367 00:23:01,880 --> 00:23:05,640 or hire workers. The whole system depends on 368 00:23:05,640 --> 00:23:09,600 participants knowing that the market is at least roughly fair. 369 00:23:10,440 --> 00:23:14,240 One of the reasons the US economy has been so successful 370 00:23:14,240 --> 00:23:18,320 over the last 90 years is that it's had some of the best 371 00:23:18,320 --> 00:23:22,640 institutions in the world, fair securities regulation, good 372 00:23:22,640 --> 00:23:26,120 consumer protections, and a functioning legal system. 373 00:23:26,360 --> 00:23:30,520 Americans invest confidently because they broadly trust the 374 00:23:30,520 --> 00:23:35,600 system, and American businesses have access to capital because 375 00:23:35,600 --> 00:23:38,000 investors are willing to put money in. 376 00:23:38,360 --> 00:23:41,560 In countries where investors know they'll be ripped off, they 377 00:23:41,560 --> 00:23:45,880 behave like the fish on the poker websites did and log off. 378 00:23:46,120 --> 00:23:49,120 This is extremely economically harmful. 379 00:23:49,480 --> 00:23:52,800 Prediction markets don't raise capital for anything. 380 00:23:53,000 --> 00:23:56,320 They don't fund new businesses or build infrastructure. 381 00:23:56,520 --> 00:24:00,040 They just move money from the pockets of retail betters into 382 00:24:00,040 --> 00:24:04,240 the pockets of quantitative algorithms and apparently people 383 00:24:04,240 --> 00:24:08,080 with high level security clearances, which is a rather 384 00:24:08,080 --> 00:24:11,560 elegant system if you think about it, just not for the 385 00:24:11,560 --> 00:24:15,000 retail betters. You might think that it doesn't 386 00:24:15,000 --> 00:24:17,080 matter. Prediction markets aren't the 387 00:24:17,080 --> 00:24:20,480 stock market. But if this market is overseen 388 00:24:20,480 --> 00:24:24,000 by the same securities and commodities regulators that 389 00:24:24,000 --> 00:24:27,960 oversee investment markets, and the public decide that it's all 390 00:24:27,960 --> 00:24:31,760 rigged, they might not just log off from prediction markets. 391 00:24:31,920 --> 00:24:35,560 They might start to wonder if the entire system is rigged too. 392 00:24:35,840 --> 00:24:40,200 And that kind of distrust, once it sets in, is very difficult to 393 00:24:40,200 --> 00:24:43,600 reverse. So if the retail better is 394 00:24:43,600 --> 00:24:47,120 structurally disadvantaged against quantitative hedge funds 395 00:24:47,120 --> 00:24:50,600 and people with high level security clearances, the obvious 396 00:24:50,600 --> 00:24:54,760 question is what happens to them when their money runs out. 397 00:24:55,280 --> 00:24:59,040 Since 2018, when the Supreme Court cleared the way for states 398 00:24:59,040 --> 00:25:02,720 to legalize sports betting, the United States has essentially 399 00:25:02,720 --> 00:25:06,520 been running a large scale experiment in what happens when 400 00:25:06,520 --> 00:25:10,160 you make it very easy for people to gamble from their phones. 401 00:25:10,520 --> 00:25:13,840 Now that we've added prediction markets to the mix, that 402 00:25:13,840 --> 00:25:18,040 experiment has now been expanded to cover politics, pop culture 403 00:25:18,040 --> 00:25:22,440 and monetary bets. You can now lose money on almost 404 00:25:22,440 --> 00:25:26,800 anything at any time of the day without even getting out of bed. 405 00:25:27,120 --> 00:25:31,280 The results of this experiment are starting to come in too, and 406 00:25:31,280 --> 00:25:34,640 they are not great. Recent academic researcher 407 00:25:34,640 --> 00:25:38,600 highlighted by The Economist shows that the introduction of 408 00:25:38,600 --> 00:25:42,480 online betting in a state is associated with a roughly 12 409 00:25:42,480 --> 00:25:46,440 point drop in average credit scores, along with higher rates 410 00:25:46,440 --> 00:25:49,560 of personal bankruptcy and loan delinquencies. 411 00:25:50,240 --> 00:25:53,800 If you make it incredibly easy for people to gamble from their 412 00:25:53,800 --> 00:25:58,320 smartphones 24 hours a day, it turns out that a rather large 413 00:25:58,320 --> 00:26:02,720 number of them will do exactly that, right up until the point 414 00:26:02,720 --> 00:26:04,640 where their credit cards are declined. 415 00:26:04,840 --> 00:26:09,120 Now, you could argue that adults in a free society should be able 416 00:26:09,120 --> 00:26:12,840 to spend their money however they like, and I mostly agree 417 00:26:12,840 --> 00:26:15,160 with that. I don't want to be told how I 418 00:26:15,160 --> 00:26:17,480 can spend my money. It's mine. 419 00:26:17,680 --> 00:26:21,200 If someone wants to bet their rent money on a basketball game 420 00:26:21,400 --> 00:26:25,320 or the exact date of a Federal Reserve rate cut, that's their 421 00:26:25,320 --> 00:26:28,200 right. The prediction market platforms 422 00:26:28,200 --> 00:26:32,040 would certainly agree with this. They describe themselves as 423 00:26:32,040 --> 00:26:35,840 neutral marketplaces facilitating price discovery, 424 00:26:36,040 --> 00:26:39,800 which is a very dignified way of describing a website where you 425 00:26:39,800 --> 00:26:42,200 can bet on whether it'll snow in April. 426 00:26:42,800 --> 00:26:45,440 The problem is that when millions of people 427 00:26:45,440 --> 00:26:49,480 simultaneously damage their personal finances, it stops 428 00:26:49,480 --> 00:26:53,200 being a private problem. Loans start to go unpaid, 429 00:26:53,360 --> 00:26:56,560 mortgages default, and eventually the costs get 430 00:26:56,560 --> 00:26:59,200 absorbed by the broader financial system. 431 00:26:59,440 --> 00:27:03,480 And the individuals who fall into bankruptcy end up relying 432 00:27:03,480 --> 00:27:08,120 on state and federal safety Nets funded by taxpayers who in many 433 00:27:08,120 --> 00:27:12,400 cases were sensible enough not to gamble their rent money on a 434 00:27:12,400 --> 00:27:15,160 basket basketball game. So when you look at the 435 00:27:15,160 --> 00:27:18,920 mechanics of the whole thing, prediction markets start to look 436 00:27:18,920 --> 00:27:22,400 less like a truth machine and more like a wealth transfer 437 00:27:22,400 --> 00:27:25,760 mechanism. The platform takes a transaction 438 00:27:25,760 --> 00:27:29,720 fee, the quantitative algorithms extract capital from retail 439 00:27:29,720 --> 00:27:34,160 betters, the insiders extract capital from everyone, and 440 00:27:34,160 --> 00:27:37,920 society picks up the tab for the bankruptcies and the unpaid 441 00:27:37,920 --> 00:27:40,760 bills. It's a wonderful business model 442 00:27:40,920 --> 00:27:43,720 for everyone except the people using it. 443 00:27:44,240 --> 00:27:48,520 Look, prediction markets are not going to destroy the American 444 00:27:48,520 --> 00:27:51,960 financial system. The stock market survived bucket 445 00:27:51,960 --> 00:27:56,880 shops in the 1920s, penny stock boiler rooms in the 1980s, and 446 00:27:56,880 --> 00:28:00,000 whatever was going on with crypto in 2021. 447 00:28:00,240 --> 00:28:04,880 It'll likely survive this too. But it's worth noticing what has 448 00:28:04,880 --> 00:28:08,120 actually been built here. A set of platforms that are 449 00:28:08,120 --> 00:28:11,720 regulated as commodity exchanges, advised by the 450 00:28:11,720 --> 00:28:15,640 president's son, populated by hedge fund algorithms and the 451 00:28:15,640 --> 00:28:20,720 occasional military insider, and marketed to 25 year olds as a 452 00:28:20,720 --> 00:28:23,160 fun way to make sports more exciting. 453 00:28:23,400 --> 00:28:27,600 You can bet on elections, wars, the weather, and the wardrobe 454 00:28:27,600 --> 00:28:31,720 choices of reggaeton artists. You just can't bet on onions 455 00:28:31,920 --> 00:28:33,880 because that would be irresponsible. 456 00:28:34,160 --> 00:28:36,480 The advocates call it a truth machine. 457 00:28:36,640 --> 00:28:39,320 The states call it an illegal sports book. 458 00:28:39,520 --> 00:28:42,200 The Quans call it a new source of alpha. 459 00:28:42,440 --> 00:28:44,880 The retail betters call it entertainment. 460 00:28:45,160 --> 00:28:49,360 The reality is probably somewhere between all of these A 461 00:28:49,360 --> 00:28:52,480 financial product that is too sophisticated to be called 462 00:28:52,480 --> 00:28:55,560 gambling and too simple to be called investing. 463 00:28:55,840 --> 00:29:00,120 Operating in a regulatory Gray zone that exists mainly because 464 00:29:00,120 --> 00:29:04,080 no one in Washington can agree on what it actually is. 465 00:29:04,440 --> 00:29:07,000 We haven't really invented a truth machine. 466 00:29:07,120 --> 00:29:10,440 We've just found a more elaborate way of losing money 467 00:29:10,600 --> 00:29:13,680 and giving it a chart. Thanks for tuning into this 468 00:29:13,680 --> 00:29:16,640 week's podcast. The podcast is entirely 469 00:29:16,640 --> 00:29:19,320 supported by viewers like you on Patreon. 470 00:29:19,480 --> 00:29:23,120 If you'd like to support it, you can sign up using the link in 471 00:29:23,120 --> 00:29:25,720 the show note. Have a great week and talk to 472 00:29:25,720 --> 00:29:27,080 you again soon. Bye.