1 00:00:00,000 --> 00:00:00,040 2 00:00:00,040 --> 00:00:02,460 The following content is provided under a Creative 3 00:00:02,460 --> 00:00:03,870 Commons license. 4 00:00:03,870 --> 00:00:06,910 Your support will help MIT OpenCourseWare continue to 5 00:00:06,910 --> 00:00:10,560 offer high-quality educational resources for free. 6 00:00:10,560 --> 00:00:13,460 To make a donation or view additional materials from 7 00:00:13,460 --> 00:00:16,180 hundreds of MIT courses, visit mitopencourseware@ocw.mit.edu. 8 00:00:16,180 --> 00:00:26,156 9 00:00:26,156 --> 00:00:29,310 PROFESSOR: OK, so what we're going to do today is continue 10 00:00:29,310 --> 00:00:33,280 our discussion of oligopoly which we started last time. 11 00:00:33,280 --> 00:00:35,460 I want to return to the example we were using last 12 00:00:35,460 --> 00:00:41,760 time to discuss the example of Cournot competition. 13 00:00:41,760 --> 00:00:44,070 We started last time by talking about general game 14 00:00:44,070 --> 00:00:46,550 theory and the prisoner's dilemma and the concept of 15 00:00:46,550 --> 00:00:48,000 Nash equilibrium. 16 00:00:48,000 --> 00:00:50,640 Then we turned to a specific example of a Cournot 17 00:00:50,640 --> 00:00:54,150 competition, the way we can specifically model two firms 18 00:00:54,150 --> 00:00:56,420 competing in a market. 19 00:00:56,420 --> 00:01:00,830 And so if you recall the example from last time, we had 20 00:01:00,830 --> 00:01:09,680 demand curve of the form p equals 339 minus q where this 21 00:01:09,680 --> 00:01:13,220 is thousands of passengers per month. 22 00:01:13,220 --> 00:01:17,430 And we had a marginal cost of \$147. 23 00:01:17,430 --> 00:01:19,610 And we have flat marginal cost, so average cost equals 24 00:01:19,610 --> 00:01:25,130 marginal cost equals 147, equals average cost. So flat 25 00:01:25,130 --> 00:01:28,070 marginal cost. 26 00:01:28,070 --> 00:01:32,185 And what we said, so therefore if the firm was a monopolist, 27 00:01:32,185 --> 00:01:36,430 if American Airlines say was a monopolist, they would set 28 00:01:36,430 --> 00:01:39,700 marginal revenue equal to marginal cost which would be 29 00:01:39,700 --> 00:01:45,630 339 minus 2q equals 147. 30 00:01:45,630 --> 00:01:49,770 And so the monopolist was going to set a quantity of 96. 31 00:01:49,770 --> 00:01:53,720 That was the monopolist quantity with the price. 32 00:01:53,720 --> 00:01:56,520 So the quantity of monopolist is 96 and the price of 33 00:01:56,520 --> 00:02:00,500 monopolist is \$243. 34 00:02:00,500 --> 00:02:04,140 So that was the monopoly case. 35 00:02:04,140 --> 00:02:07,620 Something you should be very facile with for tomorrow night 36 00:02:07,620 --> 00:02:12,410 is solving those kinds of monopoly problems. And then we 37 00:02:12,410 --> 00:02:14,910 said, OK, but what if, in fact, it recognizes that it's 38 00:02:14,910 --> 00:02:16,060 not a monopoly. 39 00:02:16,060 --> 00:02:17,160 There's another firm in the market. 40 00:02:17,160 --> 00:02:19,450 United is in the market as well. 41 00:02:19,450 --> 00:02:23,110 Well in that case it has to consider its residual demand. 42 00:02:23,110 --> 00:02:27,930 So in that case we said the qa, its residual demand, for 43 00:02:27,930 --> 00:02:31,570 American was equal to the total market q minus the 44 00:02:31,570 --> 00:02:35,270 quantity absorbed by United, qu. 45 00:02:35,270 --> 00:02:38,820 And we showed last time what this leads to graphically. 46 00:02:38,820 --> 00:02:42,870 And so we re-handed out figure 16-3 just to review. 47 00:02:42,870 --> 00:02:45,730 The notion was that each firm develops a 48 00:02:45,730 --> 00:02:47,880 best response curve. 49 00:02:47,880 --> 00:02:52,060 Each firm says, based on what the other firm's going to do, 50 00:02:52,060 --> 00:02:54,000 here is my best level of production. 51 00:02:54,000 --> 00:02:56,950 Here's my profit maximizing level of production. 52 00:02:56,950 --> 00:03:00,000 And each firm having developed a curve, there's an 53 00:03:00,000 --> 00:03:03,130 equilibrium where those curves intersect because at that 54 00:03:03,130 --> 00:03:04,530 point both firms are happy. 55 00:03:04,530 --> 00:03:07,250 You've achieved your Nash equilibrium because at that 56 00:03:07,250 --> 00:03:10,810 point both firms are satisfied with the strategy they're 57 00:03:10,810 --> 00:03:13,350 playing given what the other firm's playing. 58 00:03:13,350 --> 00:03:16,610 So that point of intersection, both firms' best response are 59 00:03:16,610 --> 00:03:19,770 consistent with each other. 60 00:03:19,770 --> 00:03:22,410 And that's we developed last time graphically. 61 00:03:22,410 --> 00:03:25,690 Intuitively, I think it follows from the same logic, 62 00:03:25,690 --> 00:03:28,160 the idea is we're playing this game and the game will only 63 00:03:28,160 --> 00:03:29,560 have a stable outcome if we're both 64 00:03:29,560 --> 00:03:30,560 satisfied with the outcome. 65 00:03:30,560 --> 00:03:32,200 If we're not both satisfied we'll continue to change our 66 00:03:32,200 --> 00:03:34,670 behavior and it won't be stable. 67 00:03:34,670 --> 00:03:36,850 And so what I want to do now is talk about how we solve for 68 00:03:36,850 --> 00:03:38,920 this mathematically. 69 00:03:38,920 --> 00:03:41,520 So let's just think about the mathematics of solving for 70 00:03:41,520 --> 00:03:44,530 Cournot equilibrium. 71 00:03:44,530 --> 00:03:48,550 So what's the mathematics now? 72 00:03:48,550 --> 00:03:52,130 What's American Airlines's residual demand function? 73 00:03:52,130 --> 00:03:59,050 Well their price, p sub a, is equal to the total demand, 74 00:03:59,050 --> 00:04:04,140 339, minus what they supply minus what United supplies. 75 00:04:04,140 --> 00:04:09,450 So the price in the market, they don't have separate 76 00:04:09,450 --> 00:04:14,350 prices, the price in the market is going to be 339 77 00:04:14,350 --> 00:04:17,730 minus what each of them supplies. 78 00:04:17,730 --> 00:04:22,510 So now American can't control q sub u, American has to take 79 00:04:22,510 --> 00:04:25,080 that as an outside given factor. 80 00:04:25,080 --> 00:04:26,950 That's what they're responding to. 81 00:04:26,950 --> 00:04:29,900 So American says, I need to optimize this with respect to 82 00:04:29,900 --> 00:04:32,160 what I can control which is q sub a. 83 00:04:32,160 --> 00:04:32,900 So how does it do it? 84 00:04:32,900 --> 00:04:35,580 Well, it computes marginal revenue. 85 00:04:35,580 --> 00:04:38,390 Well marginal revenue is p times q sub a. 86 00:04:38,390 --> 00:04:46,590 So that's 339qa minus qa squared minus quqa. 87 00:04:46,590 --> 00:04:47,840 Just multiplying through. 88 00:04:47,840 --> 00:04:52,340 89 00:04:52,340 --> 00:04:54,660 That's their revenue function. 90 00:04:54,660 --> 00:04:58,610 So marginal revenue, differentiating with respect 91 00:04:58,610 --> 00:05:03,290 to qa which is all they can control, is going to be 339 92 00:05:03,290 --> 00:05:07,920 minus 2qa minus qu. 93 00:05:07,920 --> 00:05:09,360 That's going to be their residual 94 00:05:09,360 --> 00:05:10,610 marginal revenue function. 95 00:05:10,610 --> 00:05:13,250 96 00:05:13,250 --> 00:05:17,320 And they're going to set this equal to marginal cost. So I'm 97 00:05:17,320 --> 00:05:22,550 going to set marginal revenue equal to marginal cost. So 98 00:05:22,550 --> 00:05:26,710 they're going to set this equal to 147 which is marginal 99 00:05:26,710 --> 00:05:29,140 cost. And solving. 100 00:05:29,140 --> 00:05:40,380 you get that qa equals 96 minus 1/2qu. 101 00:05:40,380 --> 00:05:42,290 qa equals 96 minus 1/2qu. 102 00:05:42,290 --> 00:05:45,440 Once again reviewing, what we're doing is we're saying we 103 00:05:45,440 --> 00:05:48,290 set up their residual demand function. 104 00:05:48,290 --> 00:05:50,980 We compute what that implies in terms of revenues. 105 00:05:50,980 --> 00:05:53,366 We differentiate with respect to their control variable, q 106 00:05:53,366 --> 00:05:56,110 sub a, to get a marginal revenue function. 107 00:05:56,110 --> 00:06:02,990 You set that equal to marginal cost. And you solve for their 108 00:06:02,990 --> 00:06:04,450 best response curve. 109 00:06:04,450 --> 00:06:06,910 Essentially you just solve mathematically for the curve 110 00:06:06,910 --> 00:06:08,160 you see here graphically. 111 00:06:08,160 --> 00:06:11,080 You go to the graph and you look at American's best 112 00:06:11,080 --> 00:06:13,070 response curve. 113 00:06:13,070 --> 00:06:14,360 That is exactly this line. 114 00:06:14,360 --> 00:06:17,970 When qu is 0, it intersects at 196. 115 00:06:17,970 --> 00:06:21,570 When qu is 192, qa is 0. 116 00:06:21,570 --> 00:06:23,610 So you can see we've just mathematically developed where 117 00:06:23,610 --> 00:06:27,010 this best response curve comes from. 118 00:06:27,010 --> 00:06:29,355 We've just solved for that mathematically through the 119 00:06:29,355 --> 00:06:33,970 similar mathematics of monopoly profit maximization. 120 00:06:33,970 --> 00:06:36,390 In this case we're essentially saying, imagine you're a 121 00:06:36,390 --> 00:06:39,300 monopoly except for the stuff produced by the other guy. 122 00:06:39,300 --> 00:06:40,140 That's the way to think about it. 123 00:06:40,140 --> 00:06:41,810 Imagine you're a monopoly except for what the other guy 124 00:06:41,810 --> 00:06:44,780 does and you get this best response function. 125 00:06:44,780 --> 00:06:47,600 Now, since we set this problem up to be symmetric-- 126 00:06:47,600 --> 00:06:51,520 that is we said the marginal cost identical for both firms, 127 00:06:51,520 --> 00:06:54,880 and the market demand applies to both firms-- 128 00:06:54,880 --> 00:06:58,240 since we set this up to be symmetrical, then we know that 129 00:06:58,240 --> 00:07:00,840 we can also write a symmetrical best response 130 00:07:00,840 --> 00:07:07,310 function for United, which is q sub u equals 96 131 00:07:07,310 --> 00:07:09,390 minus 1/2 q sub a. 132 00:07:09,390 --> 00:07:12,690 We can solve for this but there is a shortcut. 133 00:07:12,690 --> 00:07:14,820 If it's symmetrical problem they'll have symmetrical best 134 00:07:14,820 --> 00:07:16,910 response functions. 135 00:07:16,910 --> 00:07:17,980 So we can just write the symmetrical 136 00:07:17,980 --> 00:07:20,900 best response function. 137 00:07:20,900 --> 00:07:22,580 And then we can find the equilibrium. 138 00:07:22,580 --> 00:07:24,390 Remember the steps I went through last time. 139 00:07:24,390 --> 00:07:26,200 Now we have n equations and n unknowns. 140 00:07:26,200 --> 00:07:28,820 We have two equations and two unknowns. 141 00:07:28,820 --> 00:07:32,880 We can just solve by plugging one into the other. 142 00:07:32,880 --> 00:07:41,900 And we get that qa star equals qu star equals 64. 143 00:07:41,900 --> 00:07:45,000 Which is in fact where these curves intersect. 144 00:07:45,000 --> 00:07:46,950 So that's how I found the intersection of these curves 145 00:07:46,950 --> 00:07:49,760 was solving mathematically for the optimal level of 146 00:07:49,760 --> 00:07:53,020 production given what the other guy's doing. 147 00:07:53,020 --> 00:07:55,780 And the price, what's the price? 148 00:07:55,780 --> 00:08:01,820 Well the price is 339 minus q sub a minus q sub u. 149 00:08:01,820 --> 00:08:07,890 So it's 339 minus 128 or 211. 150 00:08:07,890 --> 00:08:14,980 So you end up with each firm producing 64,000 151 00:08:14,980 --> 00:08:17,930 units a price of \$211. 152 00:08:17,930 --> 00:08:20,260 And this is the Cournot equilibrium. 153 00:08:20,260 --> 00:08:22,580 Because it's on both firms' best response function. 154 00:08:22,580 --> 00:08:25,780 Both firms are maximizing profits. 155 00:08:25,780 --> 00:08:28,160 And in maximizing profits they both have chosen the same 156 00:08:28,160 --> 00:08:29,470 point to produce and so you're done. 157 00:08:29,470 --> 00:08:30,110 They're happy. 158 00:08:30,110 --> 00:08:31,750 That's the Nash equilibrium. 159 00:08:31,750 --> 00:08:32,000 Yeah. 160 00:08:32,000 --> 00:08:33,333 AUDIENCE: So in the last lecture was 161 00:08:33,333 --> 00:08:34,720 that game theory box. 162 00:08:34,720 --> 00:08:38,833 And this would be the lower right hand box, correct? 163 00:08:38,833 --> 00:08:42,617 164 00:08:42,617 --> 00:08:45,400 PROFESSOR: That's a very good point, yes. 165 00:08:45,400 --> 00:08:49,300 This is sort of like game theory but with a whole 166 00:08:49,300 --> 00:08:50,470 distribution. 167 00:08:50,470 --> 00:08:52,610 But in some sense it's like at each point you've 168 00:08:52,610 --> 00:08:53,580 reached that point. 169 00:08:53,580 --> 00:08:56,540 But I really wouldn't draw that parallel actually. 170 00:08:56,540 --> 00:08:57,790 The parallel that's important from that is 171 00:08:57,790 --> 00:08:59,900 the equilibrium concept. 172 00:08:59,900 --> 00:09:01,905 The reason we choose the prisoner's dilemma instead of 173 00:09:01,905 --> 00:09:03,460 choosing the equilibrium concept-- 174 00:09:03,460 --> 00:09:05,990 and the equilibrium concept is, you're in equilibrium when 175 00:09:05,990 --> 00:09:10,970 each firm's dominant strategy yields an equilibrium outcome. 176 00:09:10,970 --> 00:09:13,630 Here we've just taken that intuition and solved 177 00:09:13,630 --> 00:09:15,980 mathematically for each firm's dominant strategy. 178 00:09:15,980 --> 00:09:18,092 And that yields an equilibrium outcome. 179 00:09:18,092 --> 00:09:21,870 AUDIENCE: That top left box, there's no way-- 180 00:09:21,870 --> 00:09:22,850 we shouldn't think about that at all? 181 00:09:22,850 --> 00:09:24,900 PROFESSOR: No, actually it's a great segue to what I'll talk 182 00:09:24,900 --> 00:09:27,930 about next which is a cooperative equilibrium. 183 00:09:27,930 --> 00:09:30,930 So what we've solved for is we solved for a non-cooperative 184 00:09:30,930 --> 00:09:32,400 equilibrium. 185 00:09:32,400 --> 00:09:33,470 Where they're not cooperating. 186 00:09:33,470 --> 00:09:35,280 So it's like the prisoner's dilemma in the sense they end 187 00:09:35,280 --> 00:09:38,750 up in that lower corner. 188 00:09:38,750 --> 00:09:41,050 But in some sense I think that parallel, I'm not sure how 189 00:09:41,050 --> 00:09:43,350 much that helps other than thinking about the strategies 190 00:09:43,350 --> 00:09:46,560 and about thinking that that's non-cooperative. 191 00:09:46,560 --> 00:09:50,560 We want to turn to next what happens if they can cooperate. 192 00:09:50,560 --> 00:09:52,510 But before I do so, two things. 193 00:09:52,510 --> 00:09:56,910 First of all, these problems do not have to be symmetric. 194 00:09:56,910 --> 00:09:58,920 I've just solved symmetric problem often, and you'll see 195 00:09:58,920 --> 00:10:00,620 symmetric problems, that makes the math easier. 196 00:10:00,620 --> 00:10:01,740 But they don't have to be. 197 00:10:01,740 --> 00:10:04,090 American and United can be different and you could solve 198 00:10:04,090 --> 00:10:04,600 the same math. 199 00:10:04,600 --> 00:10:08,480 And in fact Perloff goes through an example like that. 200 00:10:08,480 --> 00:10:10,450 Symmetry is not guaranteed in these problems, it's just a 201 00:10:10,450 --> 00:10:12,480 feature of the way I set this up. 202 00:10:12,480 --> 00:10:15,460 So this question is a great segue to the next topic, which 203 00:10:15,460 --> 00:10:17,790 is what about cooperative equililbria. 204 00:10:17,790 --> 00:10:22,050 Can't American and United get together and form a cartel? 205 00:10:22,050 --> 00:10:24,700 206 00:10:24,700 --> 00:10:27,320 Remember we talked in the prisoner's dilemma about how 207 00:10:27,320 --> 00:10:29,390 that would be the best outcome if the two persons could just 208 00:10:29,390 --> 00:10:30,560 trust each other. 209 00:10:30,560 --> 00:10:33,490 Well here can't American and United just trust each other 210 00:10:33,490 --> 00:10:35,680 and form a cartel? 211 00:10:35,680 --> 00:10:39,080 And what they do here mathematically is pretty easy. 212 00:10:39,080 --> 00:10:42,580 They would just say, look, let's get together and pretend 213 00:10:42,580 --> 00:10:45,060 we're one firm. 214 00:10:45,060 --> 00:10:49,320 Then let's act as that one firm monopoly would and split 215 00:10:49,320 --> 00:10:51,710 the profits. 216 00:10:51,710 --> 00:10:52,750 It's the way a cartel could work. 217 00:10:52,750 --> 00:10:55,190 They could say let's get together, we'll produce as if 218 00:10:55,190 --> 00:10:57,800 we're a monopoly and let's split the profits. 219 00:10:57,800 --> 00:10:59,670 Well what would happen if they did that? 220 00:10:59,670 --> 00:11:03,910 Well we know that the total monopoly production if they 221 00:11:03,910 --> 00:11:08,780 had been a monopoly is 96. 222 00:11:08,780 --> 00:11:12,210 So that's a big Q. The quantity for the market is 96 223 00:11:12,210 --> 00:11:19,030 and the price if they're a monopoly is \$243. 224 00:11:19,030 --> 00:11:21,300 So what they did is they said, look, we'll each do 48 225 00:11:21,300 --> 00:11:28,370 flights, for a total of 96 and we'll charge \$243 and we'll 226 00:11:28,370 --> 00:11:30,030 agree to do that. 227 00:11:30,030 --> 00:11:31,020 But what would happen? 228 00:11:31,020 --> 00:11:38,340 Well their profits to each firm would be their 48 flights 229 00:11:38,340 --> 00:11:44,510 they do times the price they get minus the marginal cost. 230 00:11:44,510 --> 00:11:47,190 Marginal cost is average cost here. 231 00:11:47,190 --> 00:11:55,040 The price they get minus the marginal cost. That would mean 232 00:11:55,040 --> 00:12:01,490 per firm the profits would be \$4,608. 233 00:12:01,490 --> 00:12:04,195 So each firm would make \$4,608 in profits. 234 00:12:04,195 --> 00:12:05,070 Actually that's four million. 235 00:12:05,070 --> 00:12:06,060 This is in thousands, right. 236 00:12:06,060 --> 00:12:09,400 So \$4,608,000. 237 00:12:09,400 --> 00:12:12,300 Let's compare that to the profits they made as Cournot 238 00:12:12,300 --> 00:12:13,690 competitors. 239 00:12:13,690 --> 00:12:15,936 Once again comparing the upper box to the lower box, going 240 00:12:15,936 --> 00:12:16,770 back to the prisoner's dilemma. 241 00:12:16,770 --> 00:12:19,830 I think that is actually useful to think about. 242 00:12:19,830 --> 00:12:21,320 Let's compare that to what happened when they were 243 00:12:21,320 --> 00:12:22,460 Cournot competitors. 244 00:12:22,460 --> 00:12:26,650 When they were Cournot they each took 64 flights. 245 00:12:26,650 --> 00:12:31,260 They each did more flights when they were competing, but 246 00:12:31,260 --> 00:12:36,890 they only got to charge \$211 with that same marginal cost. 247 00:12:36,890 --> 00:12:39,840 So their profits as Cournot competitors 248 00:12:39,840 --> 00:12:48,640 were \$4,096,000 each. 249 00:12:48,640 --> 00:12:51,810 So by monopolizing and splitting the market they 250 00:12:51,810 --> 00:12:55,360 raised their profits by 12.5%, by an eighth. 251 00:12:55,360 --> 00:12:59,200 Their profits went up by an eighth, by 12.5%, which is a 252 00:12:59,200 --> 00:13:02,920 major jump in profits by getting together and forming 253 00:13:02,920 --> 00:13:03,940 this cartel. 254 00:13:03,940 --> 00:13:06,790 Just as the two prisoners were better off not ratting each 255 00:13:06,790 --> 00:13:08,680 other out, they were both better off not 256 00:13:08,680 --> 00:13:09,780 ratting each other out. 257 00:13:09,780 --> 00:13:13,070 These guys are both better off getting together and forming a 258 00:13:13,070 --> 00:13:19,380 cartel and splitting the profits, same logic. 259 00:13:19,380 --> 00:13:22,710 So why don't we see these cartels everywhere? 260 00:13:22,710 --> 00:13:24,376 Why do we have American and United, why don't 261 00:13:24,376 --> 00:13:25,600 they just do this? 262 00:13:25,600 --> 00:13:27,790 Well that's for two reasons. 263 00:13:27,790 --> 00:13:31,530 Two reasons why we don't see these cartels everywhere. 264 00:13:31,530 --> 00:13:34,745 The first reason, so why not cartels? 265 00:13:34,745 --> 00:13:40,380 266 00:13:40,380 --> 00:13:43,860 The first reason why we don't see cartels is there's a 267 00:13:43,860 --> 00:13:45,110 fundamental instability. 268 00:13:45,110 --> 00:13:49,220 269 00:13:49,220 --> 00:13:52,680 Cooperative equilibria are fundamentally unstable because 270 00:13:52,680 --> 00:13:57,200 any one party has an incentive to cheat. 271 00:13:57,200 --> 00:14:01,190 Any one party in a cartel can make money by cheating because 272 00:14:01,190 --> 00:14:03,310 they stick the other firms in the cartel. 273 00:14:03,310 --> 00:14:05,850 So let's just talk about this for a second. 274 00:14:05,850 --> 00:14:08,800 Imagine what would happen if American decided to cheat. 275 00:14:08,800 --> 00:14:10,560 Imagine they had this cartel, it's running along, they're 276 00:14:10,560 --> 00:14:14,060 each making their \$4,608,000 per month profit. 277 00:14:14,060 --> 00:14:16,020 They're all very happy. 278 00:14:16,020 --> 00:14:19,210 Now imagine American says, wait a second, what if on the 279 00:14:19,210 --> 00:14:23,530 sly I increase my number flights from 48 to 50? 280 00:14:23,530 --> 00:14:25,760 What if I just did that and did it in a way which wasn't 281 00:14:25,760 --> 00:14:27,720 immediately transparent to United? 282 00:14:27,720 --> 00:14:32,110 So American takes its q sub a which was cranking along as 48 283 00:14:32,110 --> 00:14:35,200 and it raised it to 50. 284 00:14:35,200 --> 00:14:37,010 Well what happens? 285 00:14:37,010 --> 00:14:41,480 Well if they do that the total quantity in the market instead 286 00:14:41,480 --> 00:14:45,320 of being 96 rises to 98. 287 00:14:45,320 --> 00:14:47,270 And if it rises to 98 that means the price is going to 288 00:14:47,270 --> 00:14:48,150 fall, right? 289 00:14:48,150 --> 00:14:49,650 This is not a price discriminating monopolist, 290 00:14:49,650 --> 00:14:51,740 this is a one price monopolist. So if they want to 291 00:14:51,740 --> 00:14:54,850 sell 98 flights, the price is going to have to fall. 292 00:14:54,850 --> 00:14:59,060 So the price is going to fall to \$241 from the 293 00:14:59,060 --> 00:15:02,270 current level of \$243. 294 00:15:02,270 --> 00:15:03,790 Well you might say, gee American just shot 295 00:15:03,790 --> 00:15:04,480 itself in the foot. 296 00:15:04,480 --> 00:15:07,500 Its just lowered the price by selling more flights. 297 00:15:07,500 --> 00:15:10,510 But look what happens to American's profits. 298 00:15:10,510 --> 00:15:19,020 American's profits are now 50 times 241 minus 147 which 299 00:15:19,020 --> 00:15:24,830 equals \$4,700,000 which is higher than it 300 00:15:24,830 --> 00:15:27,390 was making ib a cartel. 301 00:15:27,390 --> 00:15:32,750 American just increased its profits to more than they were 302 00:15:32,750 --> 00:15:37,970 making when they cartelized splitting the profits. 303 00:15:37,970 --> 00:15:40,280 And what's happened to United? 304 00:15:40,280 --> 00:15:42,070 So this is profits for American. 305 00:15:42,070 --> 00:15:45,840 306 00:15:45,840 --> 00:15:49,250 United's still doing 48 flights but the price has now 307 00:15:49,250 --> 00:15:53,910 fallen to \$241. 308 00:15:53,910 --> 00:16:04,930 So United's profits fall to \$4512. 309 00:16:04,930 --> 00:16:06,760 United's profits fall. 310 00:16:06,760 --> 00:16:09,520 American's profits go up. 311 00:16:09,520 --> 00:16:11,086 So basically American has stuck it 312 00:16:11,086 --> 00:16:12,336 to United by cheating. 313 00:16:12,336 --> 00:16:15,470 314 00:16:15,470 --> 00:16:17,470 It's raised its profits at the expense of United. 315 00:16:17,470 --> 00:16:18,300 And why is that? 316 00:16:18,300 --> 00:16:19,360 What's the intuition? 317 00:16:19,360 --> 00:16:21,155 Well it goes back to our monopoly intuition. 318 00:16:21,155 --> 00:16:24,010 319 00:16:24,010 --> 00:16:28,570 Because when one firm cheats in a cartel it gets all the 320 00:16:28,570 --> 00:16:32,250 benefit of the extra quantity, but only feels part of the 321 00:16:32,250 --> 00:16:34,250 poisoning effect. 322 00:16:34,250 --> 00:16:36,090 Remember when a monopolist increases its quantity. 323 00:16:36,090 --> 00:16:36,910 It's got these two effects. 324 00:16:36,910 --> 00:16:39,390 It sells more but poisons its previous sales. 325 00:16:39,390 --> 00:16:42,060 Well here when American sells two more flights, it sells 326 00:16:42,060 --> 00:16:44,840 more, it's got all that effect like a monopolist would, but 327 00:16:44,840 --> 00:16:48,270 the poisoning effect is shared with United. 328 00:16:48,270 --> 00:16:51,210 They don't feel as much pain from that price going down 329 00:16:51,210 --> 00:16:52,860 because they're sharing that pain with United. 330 00:16:52,860 --> 00:16:54,840 They get all of the benefit from quantity going up and 331 00:16:54,840 --> 00:16:57,220 only part of the pain for price coming down. 332 00:16:57,220 --> 00:16:59,516 And that's why it makes sense to cheat. 333 00:16:59,516 --> 00:17:01,340 A monopolist would never cheat itself. 334 00:17:01,340 --> 00:17:03,410 It'd make no sense for the monopolist suddenly to go from 335 00:17:03,410 --> 00:17:06,480 96 to 98 and say, ha I've cheated myself because it 336 00:17:06,480 --> 00:17:09,900 would feel the entire pain of the poisoning effect. 337 00:17:09,900 --> 00:17:10,910 But American doesn't. 338 00:17:10,910 --> 00:17:13,119 That's shared with United. 339 00:17:13,119 --> 00:17:15,520 So it actually makes sense to cheat because they get all the 340 00:17:15,520 --> 00:17:17,228 benefit of the more quantity and only part of the cost of 341 00:17:17,228 --> 00:17:18,780 the lower price. 342 00:17:18,780 --> 00:17:24,540 And that's intuitively what's going on. 343 00:17:24,540 --> 00:17:27,599 Questions about that? 344 00:17:27,599 --> 00:17:32,850 Well knowing this, clearly United says, wait a second, if 345 00:17:32,850 --> 00:17:35,180 we form a cartel American will cheat. 346 00:17:35,180 --> 00:17:37,330 So I'm going to cheat first and the 347 00:17:37,330 --> 00:17:38,120 whole thing falls apart. 348 00:17:38,120 --> 00:17:40,020 And we get back to the repeated game intuition we 349 00:17:40,020 --> 00:17:41,480 talked about in our game theory lecture. 350 00:17:41,480 --> 00:17:44,350 Where if they can commit to never cheat, if they can 351 00:17:44,350 --> 00:17:45,900 commit to punish each other if they cheat, you might be able 352 00:17:45,900 --> 00:17:48,020 to make this work in a repeated game, but only if the 353 00:17:48,020 --> 00:17:50,740 repeated game never ends. 354 00:17:50,740 --> 00:17:53,040 So if there's either not a repeated game, or it's a 355 00:17:53,040 --> 00:17:54,980 repeated game with an end, then there's going to be 356 00:17:54,980 --> 00:17:56,140 incentive for someone to cheat. 357 00:17:56,140 --> 00:17:57,710 Once there's an incentive for someone to cheat, the whole 358 00:17:57,710 --> 00:17:59,660 thing falls apart. 359 00:17:59,660 --> 00:18:03,780 And that's why cartels are fundamentally unstable. 360 00:18:03,780 --> 00:18:09,070 Now there's a second reason why we don't see cartels which 361 00:18:09,070 --> 00:18:10,830 economists would always put second. 362 00:18:10,830 --> 00:18:13,270 We always put the economic stuff first and the legal 363 00:18:13,270 --> 00:18:14,090 stuff second. 364 00:18:14,090 --> 00:18:17,250 The second is they're illegal, but we think that's less 365 00:18:17,250 --> 00:18:22,480 important as economists because we think folks are 366 00:18:22,480 --> 00:18:25,380 often smart enough to figure their way around laws. 367 00:18:25,380 --> 00:18:26,150 Technically these are illegal. 368 00:18:26,150 --> 00:18:28,330 In fact, in the late 1800s cartels were very 369 00:18:28,330 --> 00:18:29,170 common in the US. 370 00:18:29,170 --> 00:18:30,830 They were called the trusts. 371 00:18:30,830 --> 00:18:33,610 There were trusts in sugar, railroads, other areas. 372 00:18:33,610 --> 00:18:35,940 This is how these robber barons made all their money-- 373 00:18:35,940 --> 00:18:38,120 the Rockefellers and the Vanderbilts and stuff-- was by 374 00:18:38,120 --> 00:18:41,200 building monopolies. 375 00:18:41,200 --> 00:18:42,000 That's what made their money. 376 00:18:42,000 --> 00:18:45,420 And then in the early 20th century Congress passed what 377 00:18:45,420 --> 00:18:50,040 was called anti-trust laws which explicitly forbid the 378 00:18:50,040 --> 00:18:55,160 creation of cartels or these kind of trusts. 379 00:18:55,160 --> 00:18:57,160 So technically they're illegal. 380 00:18:57,160 --> 00:19:00,380 In fact, there are lots of ways, and Perloff talks about 381 00:19:00,380 --> 00:19:02,840 this in his book, there are lots of ways that firms can 382 00:19:02,840 --> 00:19:05,130 implicitly cartelize. 383 00:19:05,130 --> 00:19:10,880 There are ways that firms can say, look we'll get together 384 00:19:10,880 --> 00:19:14,790 not under the government's auspices and we'll agree that 385 00:19:14,790 --> 00:19:16,730 you raise your price and I'll do it two weeks later and look 386 00:19:16,730 --> 00:19:17,610 like we're not a cartel. 387 00:19:17,610 --> 00:19:20,820 I'm just following you, but in fact we've agreed to do it. 388 00:19:20,820 --> 00:19:28,480 So there is some implicit cartelization and basically 389 00:19:28,480 --> 00:19:30,830 that's why there's an anti-trust division exists 390 00:19:30,830 --> 00:19:31,160 that's quite large. 391 00:19:31,160 --> 00:19:34,290 The Department of Justice is basically to prosecute exactly 392 00:19:34,290 --> 00:19:35,750 those cases when firms try to do that. 393 00:19:35,750 --> 00:19:38,480 394 00:19:38,480 --> 00:19:41,730 Now in fact, however, the government has a bit of blood 395 00:19:41,730 --> 00:19:44,210 on its hands in this respect, because sometimes the 396 00:19:44,210 --> 00:19:46,390 government actually promotes cartels even though it might 397 00:19:46,390 --> 00:19:47,970 not mean to do so. 398 00:19:47,970 --> 00:19:52,480 And the best example was the 1981 voluntary export 399 00:19:52,480 --> 00:19:56,500 restraint policy, the Reagan administration. 400 00:19:56,500 --> 00:20:00,070 1981 was the worst recession before this one. 401 00:20:00,070 --> 00:20:02,085 Actually unemployment was higher than this one. 402 00:20:02,085 --> 00:20:04,060 This is a very bad recession, but actually unemployment was 403 00:20:04,060 --> 00:20:05,660 higher in 1981. 404 00:20:05,660 --> 00:20:08,430 It was well over 10%. 405 00:20:08,430 --> 00:20:11,940 And a big place where unemployment was high was in 406 00:20:11,940 --> 00:20:13,670 the manufacturing sector, particularly auto 407 00:20:13,670 --> 00:20:14,610 manufacturing. 408 00:20:14,610 --> 00:20:16,440 We were getting killed by Japan. 409 00:20:16,440 --> 00:20:18,660 Basically this was the first wave of Japan really producing 410 00:20:18,660 --> 00:20:20,490 much better cars than the US. 411 00:20:20,490 --> 00:20:22,180 And US car manufacturers were suffering 412 00:20:22,180 --> 00:20:24,060 enormously as a result. 413 00:20:24,060 --> 00:20:27,700 So basically the Reagan administration was trying to 414 00:20:27,700 --> 00:20:30,240 figure out what to do about this. 415 00:20:30,240 --> 00:20:32,070 One thing you could do, and we'll talk about in 416 00:20:32,070 --> 00:20:34,150 international trade in a few lectures, you could impose a 417 00:20:34,150 --> 00:20:37,200 quota and you could actually say to Japanese producers, you 418 00:20:37,200 --> 00:20:38,460 can't sell in America. 419 00:20:38,460 --> 00:20:39,965 But Reagan was a Republican. 420 00:20:39,965 --> 00:20:41,480 And Republicans are for free trade. 421 00:20:41,480 --> 00:20:43,730 And he really couldn't do that politically. 422 00:20:43,730 --> 00:20:45,740 Plus economists have often said that's a bad idea. 423 00:20:45,740 --> 00:20:47,960 And we'll talk about why that's a bad idea. 424 00:20:47,960 --> 00:20:51,410 What Reagan did instead was said to the Japanese auto 425 00:20:51,410 --> 00:20:53,640 manufacturers, look, I'm a charismatic guy. 426 00:20:53,640 --> 00:20:54,500 I'm Ronald Reagan. 427 00:20:54,500 --> 00:20:56,570 Why don't we get together in a room and I'll convince you to 428 00:20:56,570 --> 00:20:58,810 just export less to America. 429 00:20:58,810 --> 00:21:01,220 Let's have a voluntary export restraint agreement. 430 00:21:01,220 --> 00:21:02,990 Where I'm not going to put on a quota, because that would be 431 00:21:02,990 --> 00:21:04,980 bad, I'm going to actually suggest to you, I'm going to 432 00:21:04,980 --> 00:21:08,940 make you an offer you can't refuse that you just export 433 00:21:08,940 --> 00:21:11,530 less to the US. 434 00:21:11,530 --> 00:21:14,230 It was basically a back door way of imposing a quota 435 00:21:14,230 --> 00:21:17,880 without looking as politically bad as actually imposing one. 436 00:21:17,880 --> 00:21:21,390 And in fact, Japanese automakers agreed to this and 437 00:21:21,390 --> 00:21:24,630 reduced their sales of cars in the US. 438 00:21:24,630 --> 00:21:27,550 Now why would they agree to this? 439 00:21:27,550 --> 00:21:29,680 Now there's two explanations. 440 00:21:29,680 --> 00:21:31,850 One is they could have thought if they didn't then a quota 441 00:21:31,850 --> 00:21:32,580 was coming. 442 00:21:32,580 --> 00:21:35,320 They could have said, look, this is kind of an olive 443 00:21:35,320 --> 00:21:38,050 branch before they lower the hammer. 444 00:21:38,050 --> 00:21:39,460 That's a really bad set of metaphors, 445 00:21:39,460 --> 00:21:40,740 but you get the idea. 446 00:21:40,740 --> 00:21:43,020 That basically look, if we don't do this, then he's going 447 00:21:43,020 --> 00:21:44,840 to impose a quota. 448 00:21:44,840 --> 00:21:46,890 The other is they could have said, thank you Ronald Reagan, 449 00:21:46,890 --> 00:21:49,160 you've just formed a cartel for us. 450 00:21:49,160 --> 00:21:51,690 We can't stop from competing against each other. 451 00:21:51,690 --> 00:21:53,900 We can't stop cheating. 452 00:21:53,900 --> 00:21:55,510 We'd like to form a cartel but we can't. 453 00:21:55,510 --> 00:21:57,610 But you've told us we have to cartelize because you've 454 00:21:57,610 --> 00:21:59,680 limited how much we can sell. 455 00:21:59,680 --> 00:22:03,130 You've essentially said, look you can't sell more than this. 456 00:22:03,130 --> 00:22:04,410 You have to agree to sell more than that. 457 00:22:04,410 --> 00:22:06,470 You've essentially allowed us a framework where we can get 458 00:22:06,470 --> 00:22:09,200 together and implicitly cartelize in a way where it's 459 00:22:09,200 --> 00:22:10,820 hard to cheat because then we'd be violating the 460 00:22:10,820 --> 00:22:12,990 agreement with America. 461 00:22:12,990 --> 00:22:16,410 So basically what Reagan gave the Japanese was a cartel 462 00:22:16,410 --> 00:22:18,130 enforcing device. 463 00:22:18,130 --> 00:22:21,600 What American and United need is some way where American and 464 00:22:21,600 --> 00:22:24,540 United can monitor each other to make sure they don't cheat. 465 00:22:24,540 --> 00:22:25,820 That's what Reagan gave the Japanese. 466 00:22:25,820 --> 00:22:27,530 Said, I'm going to report on how many cars are selling in 467 00:22:27,530 --> 00:22:29,910 the US and you're going to agree to limit it to this. 468 00:22:29,910 --> 00:22:32,260 And they're like, great, that gives us the tool we need to 469 00:22:32,260 --> 00:22:34,330 enforce the cartel we wanted in the first place. 470 00:22:34,330 --> 00:22:35,480 Thank you very much. 471 00:22:35,480 --> 00:22:36,580 They made a ton of money. 472 00:22:36,580 --> 00:22:38,930 American consumers lost out. 473 00:22:38,930 --> 00:22:40,720 And I'll talk more about international trade on why 474 00:22:40,720 --> 00:22:41,500 that happens. 475 00:22:41,500 --> 00:22:43,170 But the bottom line is, this end up 476 00:22:43,170 --> 00:22:46,070 being just like a quota. 477 00:22:46,070 --> 00:22:48,335 The bad thing about quotas and things like this is that 478 00:22:48,335 --> 00:22:53,750 essentially they raise prices and hurt American consumers. 479 00:22:53,750 --> 00:22:55,420 And in this case they were bad for American consumers and 480 00:22:55,420 --> 00:22:57,340 very good for Japanese producers. 481 00:22:57,340 --> 00:23:00,710 We essentially increased profits for Japanese producers 482 00:23:00,710 --> 00:23:02,530 which they used to develop better cars and then came in 483 00:23:02,530 --> 00:23:04,810 later and killed us in the car market anyway. 484 00:23:04,810 --> 00:23:08,680 So it was dynamically a pretty terrible policy. 485 00:23:08,680 --> 00:23:14,100 So that's the legal issues around cartels. 486 00:23:14,100 --> 00:23:16,900 Now what I want to do now is step back for a second and 487 00:23:16,900 --> 00:23:22,940 say, we've now talked about three different kinds of 488 00:23:22,940 --> 00:23:27,300 market forms. We've talked about perfect competition. 489 00:23:27,300 --> 00:23:28,810 We've talked about monopoly. 490 00:23:28,810 --> 00:23:30,070 And we've talked about oligopoly. 491 00:23:30,070 --> 00:23:32,070 In particular, non-cooperative oligopoly which is a more 492 00:23:32,070 --> 00:23:33,790 interesting case since within cartels it's going to be hard 493 00:23:33,790 --> 00:23:34,950 to sustain. 494 00:23:34,950 --> 00:23:36,230 Non cooperative oligopoly. 495 00:23:36,230 --> 00:23:37,190 We've talked about three. 496 00:23:37,190 --> 00:23:38,750 How do these compare? 497 00:23:38,750 --> 00:23:41,730 So let's just do ourselves a little chart. 498 00:23:41,730 --> 00:23:42,990 Let's consider the three cases. 499 00:23:42,990 --> 00:23:51,010 We've got monopoly, oligopoly, and perfect 500 00:23:51,010 --> 00:23:54,120 competition, three cases. 501 00:23:54,120 --> 00:23:56,970 Let's ask how they compare on quantity and 502 00:23:56,970 --> 00:23:59,500 on profits per firm. 503 00:23:59,500 --> 00:24:03,750 Quantity produced in the market and profits per firm. 504 00:24:03,750 --> 00:24:06,840 And let's do it for this American, United example. 505 00:24:06,840 --> 00:24:09,940 Well we know the monopoly case, we told you that in the 506 00:24:09,940 --> 00:24:13,210 monopoly case 96 units get produced in the market. 507 00:24:13,210 --> 00:24:20,500 And the profits per firm are \$4,608,000. 508 00:24:20,500 --> 00:24:24,430 In the oligopoly case that's the Cournot case, in the 509 00:24:24,430 --> 00:24:27,920 Cournot oligopoly case we said the total amount in the market 510 00:24:27,920 --> 00:24:32,180 was 64 per firm or 128,000. 511 00:24:32,180 --> 00:24:34,010 And the profits per firm were lower. 512 00:24:34,010 --> 00:24:40,510 They were \$4,096,000 per firm. 513 00:24:40,510 --> 00:24:43,910 Now what about in perfect competition? 514 00:24:43,910 --> 00:24:46,120 What's the quantity that's going to be sold, what's the 515 00:24:46,120 --> 00:24:49,710 price going to be in perfect competition? 516 00:24:49,710 --> 00:24:50,670 We haven't done perfect competition. 517 00:24:50,670 --> 00:24:51,160 You can tell me. 518 00:24:51,160 --> 00:24:51,920 You can just do that intuitively. 519 00:24:51,920 --> 00:24:54,310 What's going to happen in this market if there's perfect 520 00:24:54,310 --> 00:24:55,090 competition what's this price going to be? 521 00:24:55,090 --> 00:24:57,090 Somebody raise their hand and tell me. 522 00:24:57,090 --> 00:24:57,870 Yeah in the back. 523 00:24:57,870 --> 00:24:59,271 AUDIENCE: [INAUDIBLE] 524 00:24:59,271 --> 00:25:01,920 PROFESSOR: 147, because price equals marginal cost in 525 00:25:01,920 --> 00:25:03,130 perfect competition. 526 00:25:03,130 --> 00:25:08,890 So quantity would be 339 minus 147 or 192. 527 00:25:08,890 --> 00:25:10,450 There will be 192 units sold. 528 00:25:10,450 --> 00:25:11,970 And what will profits be? 529 00:25:11,970 --> 00:25:14,130 Zero. 530 00:25:14,130 --> 00:25:15,420 Zero profits in perfect competition. 531 00:25:15,420 --> 00:25:17,640 With a flat marginal cost there will be zero profits 532 00:25:17,640 --> 00:25:20,250 with perfect competition. 533 00:25:20,250 --> 00:25:33,970 So we can now compare this and we could say, an oligopoly 534 00:25:33,970 --> 00:25:38,140 case leads to less output and more profits than does perfect 535 00:25:38,140 --> 00:25:39,630 competition. 536 00:25:39,630 --> 00:25:42,150 And a monopoly case leads to even less output and even more 537 00:25:42,150 --> 00:25:43,615 profits than does oligopoly. 538 00:25:43,615 --> 00:25:46,530 539 00:25:46,530 --> 00:25:51,360 And basically how do we think about welfare in this case. 540 00:25:51,360 --> 00:25:53,230 Well to think about welfare you'd have to actually draw 541 00:25:53,230 --> 00:25:58,490 the diagrams to compute producer and consumer surplus. 542 00:25:58,490 --> 00:26:03,435 What I could draw and show you is basically social welfare is 543 00:26:03,435 --> 00:26:06,336 highest in the perfect competition case, lowest in 544 00:26:06,336 --> 00:26:08,933 the monopoly case, and we talked a couple times about 545 00:26:08,933 --> 00:26:09,965 the deadweight loss of monopoly. 546 00:26:09,965 --> 00:26:10,970 This is not a price 547 00:26:10,970 --> 00:26:12,610 discriminating monopolist remember. 548 00:26:12,610 --> 00:26:15,500 This is a one price monopolist so there's deadweight loss. 549 00:26:15,500 --> 00:26:18,190 And it turns out if you write down the graph that it's in 550 00:26:18,190 --> 00:26:20,370 between the oligopoly case. 551 00:26:20,370 --> 00:26:24,050 Oligopolies cause some deadweight loss, it's worse 552 00:26:24,050 --> 00:26:26,420 than perfect competition but better then monopoly in terms 553 00:26:26,420 --> 00:26:28,440 of total welfare. 554 00:26:28,440 --> 00:26:31,670 And in fact, the bottom line is, we can actually pretty 555 00:26:31,670 --> 00:26:33,060 much tell what happens to welfare. 556 00:26:33,060 --> 00:26:34,920 A shortcut for thinking about welfare 557 00:26:34,920 --> 00:26:37,890 is to look at quantity. 558 00:26:37,890 --> 00:26:42,000 Because remember what social welfare is about, what causes 559 00:26:42,000 --> 00:26:44,730 deadweight loss, is trades that aren't made. 560 00:26:44,730 --> 00:26:48,330 561 00:26:48,330 --> 00:26:51,380 Under perfect competition every trade that has a 562 00:26:51,380 --> 00:26:54,620 positive social surpluses is made. 563 00:26:54,620 --> 00:26:57,420 So the closer the quantity produced comes to the perfect 564 00:26:57,420 --> 00:27:00,310 competition level, the smaller will be the deadweight loss, 565 00:27:00,310 --> 00:27:01,910 roughly speaking. 566 00:27:01,910 --> 00:27:04,480 So it isn't always true, and you can find bizarre 567 00:27:04,480 --> 00:27:08,850 deviations, but a good rule of thumb is that welfare, the 568 00:27:08,850 --> 00:27:10,450 distortion of perfect competition will be 569 00:27:10,450 --> 00:27:13,090 proportional to how much quantity falls relative to 570 00:27:13,090 --> 00:27:15,490 perfect competition. 571 00:27:15,490 --> 00:27:17,020 If quantity falls a lot relative to perfect 572 00:27:17,020 --> 00:27:19,200 competition, that's going to be a big loss in welfare. 573 00:27:19,200 --> 00:27:20,950 If quantity is falling just a little, it's going to be a 574 00:27:20,950 --> 00:27:21,740 small loss in welfare. 575 00:27:21,740 --> 00:27:23,500 And that goes back to that deadweight loss triangle. 576 00:27:23,500 --> 00:27:25,080 About the fact that it's smallest right around the 577 00:27:25,080 --> 00:27:27,076 perfect combination equilibrium and gets bigger as 578 00:27:27,076 --> 00:27:28,326 you move further away. 579 00:27:28,326 --> 00:27:30,420 580 00:27:30,420 --> 00:27:34,600 So basically perfect combination is the best, 581 00:27:34,600 --> 00:27:35,900 monopoly is the worst, oligopoly would be somewhere 582 00:27:35,900 --> 00:27:37,150 in between. 583 00:27:37,150 --> 00:27:43,380 584 00:27:43,380 --> 00:27:44,690 This leads to the question. 585 00:27:44,690 --> 00:27:46,180 I think you might look at this chart and ask yourself is 586 00:27:46,180 --> 00:27:49,850 well, gee, does this sort of say that the number of firms 587 00:27:49,850 --> 00:27:52,090 determines welfare? 588 00:27:52,090 --> 00:27:55,696 You've got one firm, two firms, infinite firms so can 589 00:27:55,696 --> 00:27:58,190 we make a more general statement about the role the 590 00:27:58,190 --> 00:27:58,930 number of firms. 591 00:27:58,930 --> 00:28:02,610 And it turns out the Cournot model we can. 592 00:28:02,610 --> 00:28:06,260 Basically in the Cournot model, the more firms there 593 00:28:06,260 --> 00:28:10,240 are, the closer you get to perfect competition. 594 00:28:10,240 --> 00:28:13,550 In the Cournot model as n increases you approach perfect 595 00:28:13,550 --> 00:28:16,040 competition. 596 00:28:16,040 --> 00:28:17,910 And in particular, and the book does the math in this, 597 00:28:17,910 --> 00:28:20,940 you don't need to know the math, it's more the intuition. 598 00:28:20,940 --> 00:28:24,440 But the book shows that the markup in a market, price 599 00:28:24,440 --> 00:28:28,020 minus marginal cost over price, in a Cournot market is 600 00:28:28,020 --> 00:28:31,410 equal to 1 over n times the elasticity of demand. 601 00:28:31,410 --> 00:28:36,250 602 00:28:36,250 --> 00:28:41,740 That is that we've said before that markup was inversely 603 00:28:41,740 --> 00:28:44,370 proportional to elasticity of demand. 604 00:28:44,370 --> 00:28:46,260 I might have used a different letter, I might have used nu 605 00:28:46,260 --> 00:28:47,050 for that elasticity. 606 00:28:47,050 --> 00:28:48,150 But this is elasticity of demand. 607 00:28:48,150 --> 00:28:49,380 I forget what letter I used last time. 608 00:28:49,380 --> 00:28:51,600 Same elasticity I've been using. 609 00:28:51,600 --> 00:28:55,200 So basically, we said before we had this equation, right? 610 00:28:55,200 --> 00:28:56,915 We said markup was inversely proportional 611 00:28:56,915 --> 00:28:58,290 to elasticity demand. 612 00:28:58,290 --> 00:29:00,540 The more inelastic was the good, the higher market power 613 00:29:00,540 --> 00:29:02,080 the monopolist had. 614 00:29:02,080 --> 00:29:06,140 Well here we now divide by n. 615 00:29:06,140 --> 00:29:08,680 So the point is, that for a given elasticity of demand the 616 00:29:08,680 --> 00:29:11,260 more firms that are competing in the oligopolistic setting, 617 00:29:11,260 --> 00:29:19,050 the closer and closer you're going to get to a zero mark-up 618 00:29:19,050 --> 00:29:21,750 with a Cournot equilibrium. 619 00:29:21,750 --> 00:29:24,500 So it's just this intuition you see from this table. 620 00:29:24,500 --> 00:29:26,110 This is one firm. 621 00:29:26,110 --> 00:29:27,770 Then oligopoly can imagine a number of branches. 622 00:29:27,770 --> 00:29:29,490 This is oligopoly with two firms. You should be able to 623 00:29:29,490 --> 00:29:30,000 show yourself. 624 00:29:30,000 --> 00:29:32,410 If there were three firms, which you can solve, that's 625 00:29:32,410 --> 00:29:33,600 three equations and three unknowns. 626 00:29:33,600 --> 00:29:36,700 It's the same math we just did. 627 00:29:36,700 --> 00:29:37,520 It's high school math. 628 00:29:37,520 --> 00:29:40,120 You can do this if you just put in three firms, do three 629 00:29:40,120 --> 00:29:42,040 equations and three unknowns, you get an even higher 630 00:29:42,040 --> 00:29:43,600 quantity and so on. 631 00:29:43,600 --> 00:29:45,360 So as the number of firms goes up you're going to move that 632 00:29:45,360 --> 00:29:47,590 quantity up towards the perfect competition level and 633 00:29:47,590 --> 00:29:50,260 profits will fall towards the zero that we get from perfect 634 00:29:50,260 --> 00:29:52,780 competition. 635 00:29:52,780 --> 00:29:55,650 And this is consistent with our rough notion. 636 00:29:55,650 --> 00:29:57,450 We talked about extremes, about perfect competition 637 00:29:57,450 --> 00:29:59,990 being lots of firms, the monopoly being one firm. 638 00:29:59,990 --> 00:30:02,870 But this confirms the rough intuition of the more firms, 639 00:30:02,870 --> 00:30:04,690 the more competition. 640 00:30:04,690 --> 00:30:06,480 That's a good solid intuition. 641 00:30:06,480 --> 00:30:08,980 642 00:30:08,980 --> 00:30:17,720 Now this raises the interesting question about 643 00:30:17,720 --> 00:30:21,170 should we ever allow the number of firms 644 00:30:21,170 --> 00:30:23,430 in a market to shrink. 645 00:30:23,430 --> 00:30:24,780 Should we ever allow the number of firms in 646 00:30:24,780 --> 00:30:27,360 the market to shrink? 647 00:30:27,360 --> 00:30:31,810 What we call that is we call that a merger. 648 00:30:31,810 --> 00:30:33,030 You've all heard of mergers. 649 00:30:33,030 --> 00:30:34,990 When two companies merge to form one company. 650 00:30:34,990 --> 00:30:39,020 Well by the logic I just told you that's going to be bad. 651 00:30:39,020 --> 00:30:41,320 I just said that the more firms you 652 00:30:41,320 --> 00:30:42,440 have, the higher welfare. 653 00:30:42,440 --> 00:30:43,695 The more you get towards perfect competition. 654 00:30:43,695 --> 00:30:46,400 655 00:30:46,400 --> 00:30:47,840 And yet mergers happen all the time. 656 00:30:47,840 --> 00:30:49,870 The government does not stop them. 657 00:30:49,870 --> 00:30:52,070 The government typically investigates them if they're 658 00:30:52,070 --> 00:30:54,810 big mergers, the government would 659 00:30:54,810 --> 00:30:56,100 typically investigate them. 660 00:30:56,100 --> 00:31:00,290 But by and large they often let them go through. 661 00:31:00,290 --> 00:31:02,370 Why is that? 662 00:31:02,370 --> 00:31:05,940 Why should the government not stop all mergers? 663 00:31:05,940 --> 00:31:08,370 And in particular, if anyone for extra credit, can anyone 664 00:31:08,370 --> 00:31:11,410 tell me what concept we covered which would explain 665 00:31:11,410 --> 00:31:13,190 why it might make sense to allow firms to merge. 666 00:31:13,190 --> 00:31:13,650 Yeah. 667 00:31:13,650 --> 00:31:14,940 AUDIENCE: The economy of scale. 668 00:31:14,940 --> 00:31:17,410 PROFESSOR: Economies of scale, exactly. 669 00:31:17,410 --> 00:31:18,980 That's the term I was looking for. 670 00:31:18,980 --> 00:31:22,520 It could be that you have two firms that are producing 671 00:31:22,520 --> 00:31:25,560 inefficiently relative to having combined production. 672 00:31:25,560 --> 00:31:28,100 So you've got one plant that produces the left shoes and 673 00:31:28,100 --> 00:31:31,850 one plant that produces the right shoes. 674 00:31:31,850 --> 00:31:32,920 And that's not a good example because they wouldn't really 675 00:31:32,920 --> 00:31:33,770 compete with each other. 676 00:31:33,770 --> 00:31:34,600 But you know the point. 677 00:31:34,600 --> 00:31:39,858 The point let's say you've got two firms competing with each 678 00:31:39,858 --> 00:31:42,170 other and they're both running a plant at half capacity. 679 00:31:42,170 --> 00:31:43,880 There's two separate plants running half capacity, much 680 00:31:43,880 --> 00:31:46,090 more efficient to have them merge and have one plant run 681 00:31:46,090 --> 00:31:47,920 at full capacity. 682 00:31:47,920 --> 00:31:51,020 And yes, if they merged they would then be a monopolist, 683 00:31:51,020 --> 00:31:54,440 but it's possible the cost efficiencies could be so high 684 00:31:54,440 --> 00:31:56,960 a monopolist with lower marginal cost might be better 685 00:31:56,960 --> 00:31:59,560 than oligopolist with higher marginal cost. This is kind of 686 00:31:59,560 --> 00:32:02,053 like the patent example I did last lecture 687 00:32:02,053 --> 00:32:03,140 or the lecture before. 688 00:32:03,140 --> 00:32:05,120 Where I talked about how a patent could be a good thing 689 00:32:05,120 --> 00:32:06,210 or a bad thing. 690 00:32:06,210 --> 00:32:09,370 It depends on how much demand increases due to the patent. 691 00:32:09,370 --> 00:32:10,940 It's the same notion here. 692 00:32:10,940 --> 00:32:13,920 The tradeoff is on the one hand you get marginal cost 693 00:32:13,920 --> 00:32:21,890 down by allowing a merger, on the other hand you reduce the 694 00:32:21,890 --> 00:32:24,020 number of firms in the market by allowing a merger. 695 00:32:24,020 --> 00:32:25,540 And there's that tradeoff. 696 00:32:25,540 --> 00:32:29,850 You're going to raise each firm's mark-up ability but by 697 00:32:29,850 --> 00:32:32,270 lowering marginal cost you might actually in essence, 698 00:32:32,270 --> 00:32:34,090 might in the end lower the price. 699 00:32:34,090 --> 00:32:35,490 And that's exactly what the Department of 700 00:32:35,490 --> 00:32:38,250 Justice does every day. 701 00:32:38,250 --> 00:32:39,950 And this is what is the hundreds of economists that 702 00:32:39,950 --> 00:32:41,380 work for the Department of Justice do, they have quite an 703 00:32:41,380 --> 00:32:42,410 interesting job. 704 00:32:42,410 --> 00:32:44,410 They have to go ahead and evaluate these mergers. 705 00:32:44,410 --> 00:32:46,715 People want to merge and they have to say, well which effect 706 00:32:46,715 --> 00:32:48,720 is going to be larger. 707 00:32:48,720 --> 00:32:51,490 How do we decide whether the savings from economies of 708 00:32:51,490 --> 00:32:54,900 scale exceed the cost and increase market power? 709 00:32:54,900 --> 00:32:58,790 And that's just an incredibly hard thing to do. 710 00:32:58,790 --> 00:33:02,060 That turns out to be an incredibly hard thing to do. 711 00:33:02,060 --> 00:33:04,540 And part of the reason it's hard is exactly comes back to 712 00:33:04,540 --> 00:33:06,000 what I talked about when I talked about regulating 713 00:33:06,000 --> 00:33:06,950 monopolies. 714 00:33:06,950 --> 00:33:10,090 Because the data is controlled by the market participants. 715 00:33:10,090 --> 00:33:12,560 And so you can have necessarily good data on what 716 00:33:12,560 --> 00:33:13,440 it's going to do. 717 00:33:13,440 --> 00:33:17,060 So a great example is hospitals. 718 00:33:17,060 --> 00:33:19,560 Over the last 20 years there's been enormous mergers in the 719 00:33:19,560 --> 00:33:21,040 hospital sector. 720 00:33:21,040 --> 00:33:23,300 Even here in Boston we've seen huge mergers of things like 721 00:33:23,300 --> 00:33:25,766 the Beth Israel and Deaconess hospitals merge. 722 00:33:25,766 --> 00:33:28,280 Other hospitals merged. 723 00:33:28,280 --> 00:33:30,770 These mergers were big and had to go before regulators. 724 00:33:30,770 --> 00:33:32,760 And the hospitals rightly argue, look, there's huge 725 00:33:32,760 --> 00:33:36,120 economies of scale in hospital production. 726 00:33:36,120 --> 00:33:39,050 Partly because hospitals are rarely full. 727 00:33:39,050 --> 00:33:40,880 They have to have excess capacity in case somebody 728 00:33:40,880 --> 00:33:41,480 comes in injured. 729 00:33:41,480 --> 00:33:42,890 They have to have an extra bed around. 730 00:33:42,890 --> 00:33:45,440 They don't want people sleeping in the halls. 731 00:33:45,440 --> 00:33:48,040 As a result, in a situation where you're rarely full, 732 00:33:48,040 --> 00:33:50,370 where you're constantly under capacity, there could be huge 733 00:33:50,370 --> 00:33:53,950 economies of scale from combining and using your 734 00:33:53,950 --> 00:33:56,070 combined space more effectively. 735 00:33:56,070 --> 00:33:57,710 So pretty compelling argument. 736 00:33:57,710 --> 00:34:02,220 And for that reason, pretty plausible. 737 00:34:02,220 --> 00:34:03,490 Most hospital mergers were allowed through. 738 00:34:03,490 --> 00:34:06,300 Very few hospital mergers were rejected. 739 00:34:06,300 --> 00:34:07,670 Well what happened? 740 00:34:07,670 --> 00:34:10,310 It turns out the hospitals just lied. 741 00:34:10,310 --> 00:34:12,630 They kept exactly the same production process they had 742 00:34:12,630 --> 00:34:15,929 before the merger and just raised prices. 743 00:34:15,929 --> 00:34:19,159 So in fact, Beth Israel and Deaconess each kept their 744 00:34:19,159 --> 00:34:21,520 building just like they had before the merger. 745 00:34:21,520 --> 00:34:22,980 They just charged more. 746 00:34:22,980 --> 00:34:26,050 Essentially they just cartelized. 747 00:34:26,050 --> 00:34:28,199 But the government let it go through. 748 00:34:28,199 --> 00:34:30,920 And what we've done, essentially, is we're fooled 749 00:34:30,920 --> 00:34:34,639 by the theory of these economies of scale, when in 750 00:34:34,639 --> 00:34:38,030 fact there was nothing to force the companies to 751 00:34:38,030 --> 00:34:40,719 necessarily do much to realize them. 752 00:34:40,719 --> 00:34:43,520 We just allowed them to have market power. 753 00:34:43,520 --> 00:34:46,310 So basically this is an incredibly tricky and 754 00:34:46,310 --> 00:34:49,050 difficult issue in deciding whether to allow mergers or 755 00:34:49,050 --> 00:34:52,856 not is basically evaluating whether in practice you'll see 756 00:34:52,856 --> 00:34:54,880 the economies of scale that justify the 757 00:34:54,880 --> 00:34:56,310 increased market power. 758 00:34:56,310 --> 00:35:00,590 And that's why the Department of Justice has a hard job. 759 00:35:00,590 --> 00:35:01,840 Questions about that? 760 00:35:01,840 --> 00:35:04,230 761 00:35:04,230 --> 00:35:08,560 One other point I want to make here is about the size the 762 00:35:08,560 --> 00:35:14,990 market is, I've talked about how mathematically the size of 763 00:35:14,990 --> 00:35:17,930 the market will lead to more competition. 764 00:35:17,930 --> 00:35:20,360 There's also another reason, which is the bigger the 765 00:35:20,360 --> 00:35:23,900 market, the harder to enforce a cartel. 766 00:35:23,900 --> 00:35:29,365 If there's two players in the market and one cheats, the 767 00:35:29,365 --> 00:35:32,790 other one has the first one whacked or something. 768 00:35:32,790 --> 00:35:35,630 You know what's going on, you can keep an eye on each other, 769 00:35:35,630 --> 00:35:37,580 you know who's to blame. 770 00:35:37,580 --> 00:35:39,920 When there's lots of players in the market it becomes-- 771 00:35:39,920 --> 00:35:43,070 so let's say there was 10 airlines flying, it becomes 772 00:35:43,070 --> 00:35:45,390 hard to figure out who's cheating. 773 00:35:45,390 --> 00:35:48,430 Or more still, imagine the most important cartel in the 774 00:35:48,430 --> 00:35:50,100 world which is OPEC. 775 00:35:50,100 --> 00:35:53,210 The Organization of Petroleum Exporting Countries. 776 00:35:53,210 --> 00:35:56,270 This is an organization of 20 plus countries who control the 777 00:35:56,270 --> 00:35:58,100 world's oil supply. 778 00:35:58,100 --> 00:36:00,590 But there's no good way of tracking who's 779 00:36:00,590 --> 00:36:01,370 selling what oil. 780 00:36:01,370 --> 00:36:04,160 We just know the total amount of oil sold out there. 781 00:36:04,160 --> 00:36:07,600 So if a country like Venezuela run by some nutty guy decides 782 00:36:07,600 --> 00:36:10,930 to cheat and throw a bunch of oil on the market, it's hard 783 00:36:10,930 --> 00:36:14,930 for the other members of OPEC to actually view that. 784 00:36:14,930 --> 00:36:17,280 OPEC's efficiency has varied over time. 785 00:36:17,280 --> 00:36:19,485 At times OPEC has not done such a good job of controlling 786 00:36:19,485 --> 00:36:22,730 the supply of oil in the world, at other times it has. 787 00:36:22,730 --> 00:36:24,460 But basically the more players, 788 00:36:24,460 --> 00:36:26,400 the harder it becomes. 789 00:36:26,400 --> 00:36:29,400 And as a good example, which is the cartel to produce 790 00:36:29,400 --> 00:36:32,140 mercury was an existing cartel. 791 00:36:32,140 --> 00:36:35,150 Italy and Spain essentially had controlled the market to 792 00:36:35,150 --> 00:36:38,010 sell mercury. 793 00:36:38,010 --> 00:36:39,780 They had a very well functioning cartel. 794 00:36:39,780 --> 00:36:43,640 Other countries entered and the cartel broke down. 795 00:36:43,640 --> 00:36:44,790 So we can see this. 796 00:36:44,790 --> 00:36:48,980 And this becomes a big issue in thinking about the benefits 797 00:36:48,980 --> 00:36:51,220 of enforcing competition. 798 00:36:51,220 --> 00:36:53,230 And becomes a big issue in government policy. 799 00:36:53,230 --> 00:36:54,750 Here's another government policy issue which you may 800 00:36:54,750 --> 00:36:58,210 followed lately which is the rare earths issue. 801 00:36:58,210 --> 00:37:00,170 You guys all know as scientists better than I, the 802 00:37:00,170 --> 00:37:01,680 importance of rare earths, but these are materials which are 803 00:37:01,680 --> 00:37:03,910 very important for putting in cell phone 804 00:37:03,910 --> 00:37:04,930 batteries and other things. 805 00:37:04,930 --> 00:37:07,210 These are important new materials 806 00:37:07,210 --> 00:37:08,890 for the modern economy. 807 00:37:08,890 --> 00:37:11,710 The US used to be the dominant producer of rare earths and 808 00:37:11,710 --> 00:37:13,870 then China turned out were able to produce it much more 809 00:37:13,870 --> 00:37:16,410 efficiently, so we just let China take it. 810 00:37:16,410 --> 00:37:20,560 So now 97% of the world's rare earths are produced in China. 811 00:37:20,560 --> 00:37:23,226 China's now started saying things like, Japan we're mad 812 00:37:23,226 --> 00:37:25,980 at you, we're not selling you rare earths, which is a huge, 813 00:37:25,980 --> 00:37:27,310 huge problem. 814 00:37:27,310 --> 00:37:29,200 And basically we've allowed China to essentially 815 00:37:29,200 --> 00:37:31,160 monopolize this market. 816 00:37:31,160 --> 00:37:35,590 Now we tend to think in economics it's a bad idea for 817 00:37:35,590 --> 00:37:38,330 the government to subsidize businesses. 818 00:37:38,330 --> 00:37:39,920 The government should just let the economy work and decide 819 00:37:39,920 --> 00:37:40,650 what works. 820 00:37:40,650 --> 00:37:42,000 But this may be a case where the government wants to say, 821 00:37:42,000 --> 00:37:44,220 no, we are going to actually subsidize rare earths 822 00:37:44,220 --> 00:37:46,350 production in America, to make sure on national security 823 00:37:46,350 --> 00:37:49,650 grounds we have enough in case China decides to shut us off. 824 00:37:49,650 --> 00:37:51,490 And that's exactly the kind of issue the government has to 825 00:37:51,490 --> 00:37:56,230 face in terms of trying to decide whether or not to try 826 00:37:56,230 --> 00:37:58,320 to actively get in there and promote competition versus 827 00:37:58,320 --> 00:38:01,280 just letting the market work. 828 00:38:01,280 --> 00:38:02,530 Questions about that? 829 00:38:02,530 --> 00:38:04,840 830 00:38:04,840 --> 00:38:08,660 Now I want to talk about one last thing before we stop, 831 00:38:08,660 --> 00:38:13,110 which is the fact that Cournot competition is not the only 832 00:38:13,110 --> 00:38:15,930 model of competition out there. 833 00:38:15,930 --> 00:38:17,410 In fact it might not even be the first one 834 00:38:17,410 --> 00:38:18,990 that comes to mind. 835 00:38:18,990 --> 00:38:22,120 In fact, if I said to you, imagine two firms competing to 836 00:38:22,120 --> 00:38:26,200 sell something, the first thing to come to your mind 837 00:38:26,200 --> 00:38:30,470 would probably not be best response functions. 838 00:38:30,470 --> 00:38:33,060 If you say, well they just compete over the price. 839 00:38:33,060 --> 00:38:34,670 And if there was really tough competition the price would 840 00:38:34,670 --> 00:38:35,980 just come down. 841 00:38:35,980 --> 00:38:39,640 And the limit, if they're incredibly good competitors, 842 00:38:39,640 --> 00:38:43,680 the price would have to come down close to marginal cost. 843 00:38:43,680 --> 00:38:49,250 Well that is a different model of competition that we call 844 00:38:49,250 --> 00:38:53,575 Bertrand or price competition. 845 00:38:53,575 --> 00:39:01,260 846 00:39:01,260 --> 00:39:03,670 Cournot competition is competition over quantities. 847 00:39:03,670 --> 00:39:06,350 You get together and decide how much quantity to produce 848 00:39:06,350 --> 00:39:08,210 given what the other guy is producing and then the market 849 00:39:08,210 --> 00:39:09,860 gives you the price. 850 00:39:09,860 --> 00:39:14,450 Bertrand competition is the flip of that, which is firms 851 00:39:14,450 --> 00:39:17,540 set their price and they produce whatever the market 852 00:39:17,540 --> 00:39:20,250 demands at that price. 853 00:39:20,250 --> 00:39:22,440 Once again, Cournot competition is firms set their 854 00:39:22,440 --> 00:39:26,340 quantity and then they set the price reading it 855 00:39:26,340 --> 00:39:27,940 off a demand curve. 856 00:39:27,940 --> 00:39:29,100 Bertrand competition is totally different. 857 00:39:29,100 --> 00:39:31,110 They say, we're just going to compete over price and we'll 858 00:39:31,110 --> 00:39:32,860 just produce whatever quantity you want at that price. 859 00:39:32,860 --> 00:39:36,240 We don't really even care what demand is. 860 00:39:36,240 --> 00:39:38,570 We're just going to compete over price. 861 00:39:38,570 --> 00:39:44,090 And the idea is that you have some marginal cost that firms 862 00:39:44,090 --> 00:39:46,360 will never go below in the long run. 863 00:39:46,360 --> 00:39:48,390 And the question is how much will they compete down to that 864 00:39:48,390 --> 00:39:51,640 marginal cost. In practice, what's crazy about Bertrand 865 00:39:51,640 --> 00:39:55,240 competition is it's possible with two firms you can get to 866 00:39:55,240 --> 00:39:59,070 the perfectly competitive outcome, with only two firms. 867 00:39:59,070 --> 00:40:01,010 Because those two firms are competitive enough in setting 868 00:40:01,010 --> 00:40:03,020 the price, they can fight it all the way down to the 869 00:40:03,020 --> 00:40:03,720 marginal cost. 870 00:40:03,720 --> 00:40:06,870 That is if I ever try set it to marginal cost plus \$1, 871 00:40:06,870 --> 00:40:09,200 you'll come in and set it to marginal cost plus \$0.99, 872 00:40:09,200 --> 00:40:10,970 because you can still make money. 873 00:40:10,970 --> 00:40:13,630 Now I'll come and set it to marginal cost plus \$0.98, 874 00:40:13,630 --> 00:40:14,840 because I can still make money. 875 00:40:14,840 --> 00:40:17,680 And we'll compete it all the way down until we both are 876 00:40:17,680 --> 00:40:19,810 charging marginal cost plus a penny. 877 00:40:19,810 --> 00:40:21,990 And if anyone tries to sell a higher price they can't sell 878 00:40:21,990 --> 00:40:23,140 the product because people will only buy the 879 00:40:23,140 --> 00:40:24,390 lower priced good. 880 00:40:24,390 --> 00:40:26,480 881 00:40:26,480 --> 00:40:30,930 So you could have a world with Bertrand competition where 882 00:40:30,930 --> 00:40:32,560 essentially firms compete over price. 883 00:40:32,560 --> 00:40:35,930 And where unlike this case, oh I just covered it up, unlike 884 00:40:35,930 --> 00:40:38,350 our comparison where with two firms there's still lots of 885 00:40:38,350 --> 00:40:39,730 profits to be made. 886 00:40:39,730 --> 00:40:41,920 In Bertrand competition it's possible that two firms are 887 00:40:41,920 --> 00:40:45,340 enough, just two firms are enough to get us to perfect 888 00:40:45,340 --> 00:40:46,500 competition. 889 00:40:46,500 --> 00:40:49,800 You don't need n firms. Two firms are enough. 890 00:40:49,800 --> 00:40:53,690 891 00:40:53,690 --> 00:40:54,990 And in some sense, if we thought about it, that's 892 00:40:54,990 --> 00:40:56,760 probably the first model we have in mind if we thought 893 00:40:56,760 --> 00:40:58,130 about firms competing is they compete over 894 00:40:58,130 --> 00:40:59,370 where to set the price. 895 00:40:59,370 --> 00:41:01,820 And that's essentially what the Bertrand model is. 896 00:41:01,820 --> 00:41:04,200 So two questions about that. 897 00:41:04,200 --> 00:41:08,150 The first is, well jeez, which model do we use? 898 00:41:08,150 --> 00:41:10,035 How do we know whether it's Bertrand competition or 899 00:41:10,035 --> 00:41:10,670 Cournot competition? 900 00:41:10,670 --> 00:41:12,440 How do we know which model to use? 901 00:41:12,440 --> 00:41:14,760 We've got a market with two firms in it, you tell us the 902 00:41:14,760 --> 00:41:17,310 models are going to give very different answers, how do we 903 00:41:17,310 --> 00:41:18,890 know which one to use? 904 00:41:18,890 --> 00:41:21,450 And the answer is, basically we don't know which 905 00:41:21,450 --> 00:41:22,640 one to use for sure. 906 00:41:22,640 --> 00:41:24,680 We can talk about the conditions under which 907 00:41:24,680 --> 00:41:27,020 quantity competition is more likely and under which price 908 00:41:27,020 --> 00:41:29,630 competition is more likely. 909 00:41:29,630 --> 00:41:32,540 Basically quantity competition is going to be more likely 910 00:41:32,540 --> 00:41:36,000 when there's lags in the production process. 911 00:41:36,000 --> 00:41:38,380 So that you can't flexibly say, here's my price, I'll 912 00:41:38,380 --> 00:41:40,340 produce as much as I want. 913 00:41:40,340 --> 00:41:42,995 If an airline said, I'm charging this, I'll fly as 914 00:41:42,995 --> 00:41:44,460 many guys as I want, you can't do that. 915 00:41:44,460 --> 00:41:46,600 You only have so many planes. 916 00:41:46,600 --> 00:41:48,530 Airlines can't just set a price and then meet whatever 917 00:41:48,530 --> 00:41:49,590 demand comes out. 918 00:41:49,590 --> 00:41:51,220 Airlines have to actually sent a quantity. 919 00:41:51,220 --> 00:41:55,040 They have to say there's this many flights we can do. 920 00:41:55,040 --> 00:41:58,350 So when there's lags in the production process it's going 921 00:41:58,350 --> 00:42:00,250 to be hard to do pure price competition. 922 00:42:00,250 --> 00:42:03,320 It's going to be more like quantity competition. 923 00:42:03,320 --> 00:42:07,920 That's going to be like airlines or cars. 924 00:42:07,920 --> 00:42:10,850 On the other hand, take selling cereal at a 925 00:42:10,850 --> 00:42:13,390 supermarket. 926 00:42:13,390 --> 00:42:16,260 The supermarket if it runs out of cereal the next day it can 927 00:42:16,260 --> 00:42:18,850 get a whole bunch more. 928 00:42:18,850 --> 00:42:21,060 The supermarket can almost instantaneously replenish any 929 00:42:21,060 --> 00:42:22,640 supply it's out of. 930 00:42:22,640 --> 00:42:25,210 So there, you're more likely to see same price competition 931 00:42:25,210 --> 00:42:27,680 between say cereals in the supermarket. 932 00:42:27,680 --> 00:42:30,810 People can just compare very easily. 933 00:42:30,810 --> 00:42:33,320 And if one's cheaper they'll just buy it up and if it runs 934 00:42:33,320 --> 00:42:34,640 out the supermarket will just reload the next 935 00:42:34,640 --> 00:42:36,190 day with that cereal. 936 00:42:36,190 --> 00:42:40,360 Essentially it's pretty easy to meet quantity. 937 00:42:40,360 --> 00:42:43,080 So this is not a very clean distinction. 938 00:42:43,080 --> 00:42:45,144 But the bottom line is quantity competition is going 939 00:42:45,144 --> 00:42:46,960 to be more likely when it's a hard good 940 00:42:46,960 --> 00:42:49,820 to produce on demand. 941 00:42:49,820 --> 00:42:51,885 And price competition more likely when it's an easy good 942 00:42:51,885 --> 00:42:53,510 to do that. 943 00:42:53,510 --> 00:42:55,600 That's how you're going to think about which kind of 944 00:42:55,600 --> 00:42:58,990 competition makes the most sense in which context. 945 00:42:58,990 --> 00:43:02,140 That's point one. 946 00:43:02,140 --> 00:43:04,220 Point two, and the last point I want to make. 947 00:43:04,220 --> 00:43:08,120 What can firms do to avoid Bertrand competition. 948 00:43:08,120 --> 00:43:11,140 Let's say, for example, you're a cereal producer. 949 00:43:11,140 --> 00:43:12,410 You're in a Bertrand market. 950 00:43:12,410 --> 00:43:14,340 Well this is a shitty market to be in. 951 00:43:14,340 --> 00:43:17,930 I just said two firms can drive your profits to zero. 952 00:43:17,930 --> 00:43:19,540 What do you do? 953 00:43:19,540 --> 00:43:22,161 What do cereal producers do? 954 00:43:22,161 --> 00:43:22,628 Yeah. 955 00:43:22,628 --> 00:43:23,562 AUDIENCE: Differentiate. 956 00:43:23,562 --> 00:43:26,240 PROFESSOR: They product differentiate. 957 00:43:26,240 --> 00:43:29,570 So the way to beat Bertrand competition is through product 958 00:43:29,570 --> 00:43:30,820 differentiation. 959 00:43:30,820 --> 00:43:42,300 960 00:43:42,300 --> 00:43:44,610 Once your products are different, then you can price 961 00:43:44,610 --> 00:43:46,340 above marginal cost. Then they're not perfect 962 00:43:46,340 --> 00:43:48,860 substitutes anymore and you can price above marginal cost. 963 00:43:48,860 --> 00:43:51,070 Have you ever asked yourself why the hell there's so many 964 00:43:51,070 --> 00:43:51,770 kinds of cereal. 965 00:43:51,770 --> 00:43:54,130 I mean have you been to the supermarket? 966 00:43:54,130 --> 00:43:55,470 It's crazy. 967 00:43:55,470 --> 00:43:57,410 Why are there so many kinds of toothpaste and shampoo. 968 00:43:57,410 --> 00:43:59,580 It's all the same crap. 969 00:43:59,580 --> 00:44:00,140 Why? 970 00:44:00,140 --> 00:44:01,960 Because it's product differentiation. 971 00:44:01,960 --> 00:44:05,130 It's to avoid Bertrand competition. 972 00:44:05,130 --> 00:44:07,280 Because once you can differentiate, 973 00:44:07,280 --> 00:44:08,590 you get market power. 974 00:44:08,590 --> 00:44:11,160 Differentiation provides market power. 975 00:44:11,160 --> 00:44:13,225 In the book, this is too detailed for the course, but 976 00:44:13,225 --> 00:44:15,290 in the book Perloff talks about this as monopolistic 977 00:44:15,290 --> 00:44:16,520 competition. 978 00:44:16,520 --> 00:44:18,290 Which is sort of an odd oxymoron, right? 979 00:44:18,290 --> 00:44:19,720 Monopolistic competition. 980 00:44:19,720 --> 00:44:25,840 But these are models of differentiated products where 981 00:44:25,840 --> 00:44:28,700 essentially you can say, look I'm different enough that 982 00:44:28,700 --> 00:44:31,510 essentially I can charge a higher price than my 983 00:44:31,510 --> 00:44:33,420 competitor. 984 00:44:33,420 --> 00:44:35,250 This comes back the concept of contestable 985 00:44:35,250 --> 00:44:37,150 markets we talked about. 986 00:44:37,150 --> 00:44:41,786 The idea is that, look, if you're different you can 987 00:44:41,786 --> 00:44:42,650 charge a little bit higher price. 988 00:44:42,650 --> 00:44:44,520 Once it's too high people are going to switch back to the 989 00:44:44,520 --> 00:44:45,520 other food. 990 00:44:45,520 --> 00:44:47,150 But if you make it different enough, 991 00:44:47,150 --> 00:44:49,070 people will pay something. 992 00:44:49,070 --> 00:44:53,010 The most famous example is Apple Cinnamon Cheerios. 993 00:44:53,010 --> 00:44:55,180 Cheerios have been around forever. 994 00:44:55,180 --> 00:44:57,280 And what happened was people bought 995 00:44:57,280 --> 00:44:57,840 Cheerios, it was very popular. 996 00:44:57,840 --> 00:45:00,860 But then supermarkets started producing generic Cheerios. 997 00:45:00,860 --> 00:45:03,760 And Cheerios are like pretty freaking easy to produce. 998 00:45:03,760 --> 00:45:05,480 And it turned out they were identical. 999 00:45:05,480 --> 00:45:08,020 And supermarkets charged like half as much 1000 00:45:08,020 --> 00:45:09,820 for the generic Cheerios. 1001 00:45:09,820 --> 00:45:12,130 And Kellogg's, no Kellogg's? 1002 00:45:12,130 --> 00:45:13,610 Not Kellogg's, the other one. 1003 00:45:13,610 --> 00:45:14,470 General Mills. 1004 00:45:14,470 --> 00:45:16,755 General Mills was all like, any of 1005 00:45:16,755 --> 00:45:18,590 you guys from Minnesota? 1006 00:45:18,590 --> 00:45:20,560 Mall of America has this cool General Mills thing in 1007 00:45:20,560 --> 00:45:22,310 Minnesota, I remember it. 1008 00:45:22,310 --> 00:45:25,620 So General Mills said, well gee, we're 1009 00:45:25,620 --> 00:45:26,330 getting killed here. 1010 00:45:26,330 --> 00:45:28,560 These generic Cheerios are going to wipe out our profits. 1011 00:45:28,560 --> 00:45:31,680 So they invented the Apple Cinnamon Cheerios. 1012 00:45:31,680 --> 00:45:34,370 And this was a product differentiation where 1013 00:45:34,370 --> 00:45:36,390 basically they could still charge a high price for Apple 1014 00:45:36,390 --> 00:45:38,050 Cinnamon Cheerios because it was different enough from the 1015 00:45:38,050 --> 00:45:42,120 generic cheerios the supermarket was making. 1016 00:45:42,120 --> 00:45:43,500 And they could patent them. 1017 00:45:43,500 --> 00:45:45,030 They could say, hey, we've got a different thing we can 1018 00:45:45,030 --> 00:45:48,370 patent our secret Apple Cinnamon formula. 1019 00:45:48,370 --> 00:45:50,910 So supermarkets now just can't copy us. 1020 00:45:50,910 --> 00:45:53,890 So for a while we get price greater than the marginal 1021 00:45:53,890 --> 00:45:56,200 cost. Or even if supermarkets can copy them, they at least 1022 00:45:56,200 --> 00:45:58,150 get for a while price above marginal cost until 1023 00:45:58,150 --> 00:46:01,220 supermarkets catch on and make the product. 1024 00:46:01,220 --> 00:46:04,630 So product differentiation is the way that firms fight 1025 00:46:04,630 --> 00:46:06,350 Bertrand price competition. 1026 00:46:06,350 --> 00:46:07,900 So is this bad? 1027 00:46:07,900 --> 00:46:09,350 Once again we don't know. 1028 00:46:09,350 --> 00:46:12,190 It depends how much you like Apple Cinnamon Cheerios. 1029 00:46:12,190 --> 00:46:15,970 Just like patents, it depends on whether the new good they 1030 00:46:15,970 --> 00:46:19,960 invented delivers sufficient extra consumer surplus that 1031 00:46:19,960 --> 00:46:22,670 it's worth the fact that they can market prices on it. 1032 00:46:22,670 --> 00:46:25,380 In fact, one of the most famous studies in this field 1033 00:46:25,380 --> 00:46:28,680 was done by my colleague Jerry Hausman, who actually measured 1034 00:46:28,680 --> 00:46:31,860 the demand curve for Apple Cinnamon Cheerios and the cost 1035 00:46:31,860 --> 00:46:35,040 of producing Apple Cinnamon Cheerios and found that the 1036 00:46:35,040 --> 00:46:38,070 introduction of Apple Cinnamon Cheerios raised consumer 1037 00:46:38,070 --> 00:46:40,710 welfare by \$67 million a year. 1038 00:46:40,710 --> 00:46:42,980 That even though the prices were high because of product 1039 00:46:42,980 --> 00:46:45,210 differentiation, people were so happy with this wonderful 1040 00:46:45,210 --> 00:46:49,550 new taste, it shifted demand out so much that actually 1041 00:46:49,550 --> 00:46:51,280 consumer welfare went up. 1042 00:46:51,280 --> 00:46:54,320 Even though this was a tool introduced by the companies to 1043 00:46:54,320 --> 00:46:55,990 avoid price competition, which should make 1044 00:46:55,990 --> 00:46:58,260 consumers better off. 1045 00:46:58,260 --> 00:47:04,170 So it's all about these trade-offs between how much 1046 00:47:04,170 --> 00:47:06,430 does it make consumers better off versus how much more 1047 00:47:06,430 --> 00:47:08,170 market power does it give the producers. 1048 00:47:08,170 --> 00:47:11,910 And product differentiation is another example of that. 1049 00:47:11,910 --> 00:47:15,260 Why don't we stop there and we'll come back. 1050 00:47:15,260 --> 00:47:18,750 Good luck tomorrow night and we'll come back on Wednesday 1051 00:47:18,750 --> 00:47:20,780 and move on to factor markets. 1052 00:47:20,780 --> 00:47:22,030