[Script Info]
Title: 
[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
Dialogue: 0,0:00:06.09,0:00:07.58,Default,,0000,0000,0000,,>> Let's talk a little\Nbit about game theory.
Dialogue: 0,0:00:08.41,0:00:10.90,Default,,0000,0000,0000,,Some times in economics\Npeople want to be able
Dialogue: 0,0:00:10.90,0:00:12.71,Default,,0000,0000,0000,,to describe situations
Dialogue: 0,0:00:12.71,0:00:15.10,Default,,0000,0000,0000,,that involve what we call\Nstrategic interaction.
Dialogue: 0,0:00:15.94,0:00:17.81,Default,,0000,0000,0000,,Strategic interaction just means
Dialogue: 0,0:00:17.81,0:00:22.08,Default,,0000,0000,0000,,that now not only does your pay\Noff, your profit, your utility,
Dialogue: 0,0:00:22.08,0:00:25.86,Default,,0000,0000,0000,,how ever you want to think about\Nit depend on your own choices
Dialogue: 0,0:00:26.29,0:00:29.38,Default,,0000,0000,0000,,but it also depends on the\Nchoices of other people
Dialogue: 0,0:00:29.38,0:00:32.92,Default,,0000,0000,0000,,in your market, in your\Nindustry and so on and so forth.
Dialogue: 0,0:00:33.100,0:00:37.10,Default,,0000,0000,0000,,Typical examples of strategic\Ninteraction usually involves
Dialogue: 0,0:00:37.10,0:00:39.49,Default,,0000,0000,0000,,decision among firms\Nregarding whether
Dialogue: 0,0:00:39.49,0:00:42.25,Default,,0000,0000,0000,,to cooperate or to compete.
Dialogue: 0,0:00:42.91,0:00:44.36,Default,,0000,0000,0000,,We're going to go\Nover and example
Dialogue: 0,0:00:44.36,0:00:46.40,Default,,0000,0000,0000,,that has a slightly\Ndifferent context known
Dialogue: 0,0:00:46.40,0:00:49.70,Default,,0000,0000,0000,,as the prisoner's dilemma where\Npeople are deciding whether
Dialogue: 0,0:00:49.70,0:00:52.34,Default,,0000,0000,0000,,or not to confess to\Na particular crime.
Dialogue: 0,0:00:52.48,0:00:56.48,Default,,0000,0000,0000,,The set up of the prisoner's\Ndilemma is a tad bit contrived
Dialogue: 0,0:00:56.58,0:00:57.56,Default,,0000,0000,0000,,but it goes as follows.
Dialogue: 0,0:00:58.46,0:01:01.38,Default,,0000,0000,0000,,Imagine a situation in\Nwhich two people are brought
Dialogue: 0,0:01:01.58,0:01:03.24,Default,,0000,0000,0000,,in for supposedly\Ncommitting a crime.
Dialogue: 0,0:01:04.07,0:01:06.31,Default,,0000,0000,0000,,Now these two people are\Nheld in separate cells
Dialogue: 0,0:01:06.31,0:01:08.51,Default,,0000,0000,0000,,so they can't talk to\Neach other and even
Dialogue: 0,0:01:08.51,0:01:11.32,Default,,0000,0000,0000,,if they could they couldn't\Nsomehow contract on whether
Dialogue: 0,0:01:11.32,0:01:12.72,Default,,0000,0000,0000,,or not they were going\Nto confess to the crime.
Dialogue: 0,0:01:14.24,0:01:16.36,Default,,0000,0000,0000,,The people are then\Nbrought in individually
Dialogue: 0,0:01:16.51,0:01:19.33,Default,,0000,0000,0000,,and asked do you confess\Nor do you not confess?
Dialogue: 0,0:01:20.66,0:01:23.61,Default,,0000,0000,0000,,We can represent the pay off's\Nto that sort of situation
Dialogue: 0,0:01:24.30,0:01:25.53,Default,,0000,0000,0000,,in a table as follows.
Dialogue: 0,0:01:25.92,0:01:31.46,Default,,0000,0000,0000,,You'll notice here that we\Nhave player 1 and player 2.
Dialogue: 0,0:01:31.46,0:01:33.66,Default,,0000,0000,0000,,I made things nicely color coded
Dialogue: 0,0:01:33.66,0:01:37.76,Default,,0000,0000,0000,,such that we have player 1's\Npay off's in terms of utility
Dialogue: 0,0:01:38.80,0:01:43.14,Default,,0000,0000,0000,,in blue to match play 1 and\Nplayer 2's pay off's in terms
Dialogue: 0,0:01:43.14,0:01:45.59,Default,,0000,0000,0000,,of utility in green here.
Dialogue: 0,0:01:47.11,0:01:51.39,Default,,0000,0000,0000,,So you'll notice that if neither\Nplayer confesses they just sit
Dialogue: 0,0:01:51.39,0:01:54.62,Default,,0000,0000,0000,,there and hold tight, they\Neach get a pay off of 10.
Dialogue: 0,0:01:55.90,0:01:59.57,Default,,0000,0000,0000,,If the first guy keeps quiet\Nand the second guy rats him
Dialogue: 0,0:01:59.57,0:02:05.22,Default,,0000,0000,0000,,out the second guy gets 15 while\Nthe first player gets nothing.
Dialogue: 0,0:02:06.55,0:02:10.76,Default,,0000,0000,0000,,The opposite happens here\Nif the first player rats
Dialogue: 0,0:02:10.76,0:02:13.76,Default,,0000,0000,0000,,out the second one, now\Nthe first player gets 15
Dialogue: 0,0:02:14.16,0:02:15.52,Default,,0000,0000,0000,,and the second player\Ngets nothing.
Dialogue: 0,0:02:16.57,0:02:19.56,Default,,0000,0000,0000,,And if they both try to rat\Neach other out, they both end
Dialogue: 0,0:02:19.56,0:02:24.16,Default,,0000,0000,0000,,up with 5 meaning they're better\Noff than if they just sat here
Dialogue: 0,0:02:24.16,0:02:27.07,Default,,0000,0000,0000,,and had the other guy\Nrat him out but not quite
Dialogue: 0,0:02:27.07,0:02:30.79,Default,,0000,0000,0000,,as well off collectively\Nas if they both kept quiet.
Dialogue: 0,0:02:31.63,0:02:38.24,Default,,0000,0000,0000,,The question then becomes given\Nthis structure what's going
Dialogue: 0,0:02:38.24,0:02:38.61,Default,,0000,0000,0000,,to happen.
Dialogue: 0,0:02:38.61,0:02:40.75,Default,,0000,0000,0000,,In reality both players are\Nmaking the decision of whether
Dialogue: 0,0:02:40.86,0:02:44.100,Default,,0000,0000,0000,,or not to confess at the same\Ntime but let's just pretend
Dialogue: 0,0:02:44.100,0:02:46.82,Default,,0000,0000,0000,,that they can guess or somehow\Nknow what the other person is
Dialogue: 0,0:02:46.82,0:02:50.38,Default,,0000,0000,0000,,going to do and we can ask a\Nnumber of hypothetical questions
Dialogue: 0,0:02:50.71,0:02:54.35,Default,,0000,0000,0000,,as to what the best response\Nis for these players would be.
Dialogue: 0,0:02:54.35,0:02:57.21,Default,,0000,0000,0000,,So let's take the\Nfirst case here,
Dialogue: 0,0:02:58.12,0:03:02.26,Default,,0000,0000,0000,,say if player 1 confesses\Nwhat should player 2 do?
Dialogue: 0,0:03:02.26,0:03:04.100,Default,,0000,0000,0000,,In other words what's\Nplayer 2's best response?
Dialogue: 0,0:03:06.37,0:03:07.57,Default,,0000,0000,0000,,Well, we can go over here,
Dialogue: 0,0:03:08.29,0:03:11.75,Default,,0000,0000,0000,,we say if player 1\Nconfesses we're somewhere
Dialogue: 0,0:03:11.86,0:03:18.29,Default,,0000,0000,0000,,in the bottom here and player 2\Ncan either get zero by holding
Dialogue: 0,0:03:18.29,0:03:22.46,Default,,0000,0000,0000,,out and being quiet or he\Ncan get 5 by confessing also.
Dialogue: 0,0:03:23.51,0:03:28.52,Default,,0000,0000,0000,,Five is strictly better than\Nzero so if player 1 confesses,
Dialogue: 0,0:03:28.57,0:03:36.75,Default,,0000,0000,0000,,player 2 also wants to confess.
Dialogue: 0,0:03:37.09,0:03:39.12,Default,,0000,0000,0000,,Now what about if\Nplayer 1 doesn't confess,
Dialogue: 0,0:03:40.07,0:03:44.56,Default,,0000,0000,0000,,well if player 1 doesn't\Nconfess we're up here
Dialogue: 0,0:03:44.77,0:03:48.20,Default,,0000,0000,0000,,so player 2 again has two\Noptions, he can get 10
Dialogue: 0,0:03:48.20,0:03:53.62,Default,,0000,0000,0000,,by keeping quiet or he can get\N15 by ratting out his buddy.
Dialogue: 0,0:03:54.67,0:03:59.15,Default,,0000,0000,0000,,So 15 is better than 10 so\Nif player 1 doesn't confess,
Dialogue: 0,0:04:00.78,0:04:08.77,Default,,0000,0000,0000,,player 2 still should confess.
Dialogue: 0,0:04:08.89,0:04:13.89,Default,,0000,0000,0000,,Notice here that's interesting\Nthat player 2 his best option is
Dialogue: 0,0:04:13.89,0:04:17.74,Default,,0000,0000,0000,,to confess regardless\Nof what player one does
Dialogue: 0,0:04:18.33,0:04:21.26,Default,,0000,0000,0000,,or alternatively put\Nplayer 2's best option is
Dialogue: 0,0:04:21.33,0:04:25.02,Default,,0000,0000,0000,,to confess regardless of what he\Nthinks player 1 is going to do.
Dialogue: 0,0:04:26.26,0:04:30.34,Default,,0000,0000,0000,,This type of situation is\Ncalled a dominant strategy
Dialogue: 0,0:04:30.74,0:04:33.13,Default,,0000,0000,0000,,in that confess is\Na dominant strategy
Dialogue: 0,0:04:33.19,0:04:35.89,Default,,0000,0000,0000,,for player 2 meaning it's\Nalways the best regardless
Dialogue: 0,0:04:35.89,0:04:36.83,Default,,0000,0000,0000,,of what the other guy does.
Dialogue: 0,0:04:37.98,0:04:40.97,Default,,0000,0000,0000,,Think about this the other way\Naround, say we make some guesses
Dialogue: 0,0:04:40.97,0:04:44.45,Default,,0000,0000,0000,,as to what player 2 is going\Nto do and then when we say
Dialogue: 0,0:04:44.45,0:04:47.41,Default,,0000,0000,0000,,in each case what's player 1's\Nbest response in that situation.
Dialogue: 0,0:04:49.14,0:04:50.93,Default,,0000,0000,0000,,So if player 2 confesses,
Dialogue: 0,0:04:51.53,0:04:54.92,Default,,0000,0000,0000,,what's the best thing\Nfor player 1 to do?
Dialogue: 0,0:04:55.17,0:04:57.73,Default,,0000,0000,0000,,Say if player 2 confesses\Nwe're over here
Dialogue: 0,0:04:57.73,0:05:02.65,Default,,0000,0000,0000,,on the right somewhere we\Nsay player 1 can either get 5
Dialogue: 0,0:05:02.100,0:05:07.89,Default,,0000,0000,0000,,by confessing or 0 for being\Nquiet this problem is looking
Dialogue: 0,0:05:07.89,0:05:11.13,Default,,0000,0000,0000,,strangely familiar, say\Nwell 5 is better than 0
Dialogue: 0,0:05:11.38,0:05:16.97,Default,,0000,0000,0000,,so player 1 is going\Nto want to confess.
Dialogue: 0,0:05:18.70,0:05:22.24,Default,,0000,0000,0000,,Now if player 2 doesn't\Nconfess what should player 1 do?
Dialogue: 0,0:05:23.97,0:05:26.92,Default,,0000,0000,0000,,So if player 2 doesn't\Nconfess, we're over here
Dialogue: 0,0:05:26.92,0:05:30.74,Default,,0000,0000,0000,,on the left somewhere and\Nplayer 1 can either get 10
Dialogue: 0,0:05:30.99,0:05:34.01,Default,,0000,0000,0000,,by being quiet or 15 by\Nratting out his buddy,
Dialogue: 0,0:05:34.52,0:05:44.84,Default,,0000,0000,0000,,well 15 is greater than 10 so\Nhe's going to want to confess.
Dialogue: 0,0:05:44.97,0:05:46.94,Default,,0000,0000,0000,,Notice here that\Nbecause we confessed
Dialogue: 0,0:05:47.08,0:05:51.63,Default,,0000,0000,0000,,in both cases confessing\Nis also a dominant strategy
Dialogue: 0,0:05:52.07,0:05:53.69,Default,,0000,0000,0000,,for player 1.
Dialogue: 0,0:05:54.99,0:05:58.84,Default,,0000,0000,0000,,So here I've circled player\N2's best responses in green
Dialogue: 0,0:05:59.32,0:06:02.100,Default,,0000,0000,0000,,and I've circled player\N1's best responses in blue
Dialogue: 0,0:06:04.13,0:06:06.91,Default,,0000,0000,0000,,and you'll notice there's one\Nplace here where they over lap
Dialogue: 0,0:06:07.61,0:06:11.89,Default,,0000,0000,0000,,to say that in this situation\Nwhere both parties confess both
Dialogue: 0,0:06:11.89,0:06:15.18,Default,,0000,0000,0000,,of them are responding\Nas best they can
Dialogue: 0,0:06:15.79,0:06:18.51,Default,,0000,0000,0000,,to what they think the other\Nperson is going to be doing.
Dialogue: 0,0:06:19.74,0:06:22.54,Default,,0000,0000,0000,,We say that this situation\Nhere is what's called a Nash
Dialogue: 0,0:06:22.73,0:06:27.66,Default,,0000,0000,0000,,equilibrium; more formally put a\NNash equilibrium is a situation
Dialogue: 0,0:06:27.66,0:06:31.45,Default,,0000,0000,0000,,where each player's\Naction is the best response
Dialogue: 0,0:06:31.89,0:06:34.51,Default,,0000,0000,0000,,to the other player's actions.
Dialogue: 0,0:06:35.26,0:06:38.49,Default,,0000,0000,0000,,In a situation where the players\Nare all moving simultaneously
Dialogue: 0,0:06:38.76,0:06:41.93,Default,,0000,0000,0000,,this basically means that\Neach player is reacting best
Dialogue: 0,0:06:42.10,0:06:44.28,Default,,0000,0000,0000,,to what they think the\Nother person is going to do
Dialogue: 0,0:06:44.75,0:06:46.42,Default,,0000,0000,0000,,and they're actually\Nright in their guess
Dialogue: 0,0:06:46.42,0:06:48.48,Default,,0000,0000,0000,,of what the other\Nperson is going to do.
Dialogue: 0,0:06:55.52,0:07:01.100,Default,,0000,0000,0000,,[ Pause ]
Dialogue: 0,0:07:02.50,0:07:06.80,Default,,0000,0000,0000,,Notice here that the equilibrium\Noutcome actually...it doesn't
Dialogue: 0,0:07:06.80,0:07:09.74,Default,,0000,0000,0000,,look as good as it could\Nbecause here we're saying
Dialogue: 0,0:07:09.74,0:07:12.56,Default,,0000,0000,0000,,that any equilibrium when\Npeople are acting according
Dialogue: 0,0:07:12.56,0:07:15.56,Default,,0000,0000,0000,,to their own best interest each\Nof them ends up with a payout
Dialogue: 0,0:07:15.56,0:07:19.40,Default,,0000,0000,0000,,of 5 where as if they only\Ncooperated they would each get a
Dialogue: 0,0:07:19.40,0:07:21.71,Default,,0000,0000,0000,,payout of 10.
Dialogue: 0,0:07:21.79,0:07:26.79,Default,,0000,0000,0000,,We can say here that there can\Nbe a perato improvement going
Dialogue: 0,0:07:26.79,0:07:30.84,Default,,0000,0000,0000,,from both parties confessing\Nto both parties staying quiet
Dialogue: 0,0:07:31.38,0:07:33.59,Default,,0000,0000,0000,,in that both parties\Nwould be made better off
Dialogue: 0,0:07:34.41,0:07:36.29,Default,,0000,0000,0000,,and nobody would\Nbe made worse off.
Dialogue: 0,0:07:37.67,0:07:39.84,Default,,0000,0000,0000,,Unfortunately, due to\Nthe competitive nature
Dialogue: 0,0:07:39.84,0:07:42.74,Default,,0000,0000,0000,,of the this game that's\Nnot what's going to result
Dialogue: 0,0:07:42.74,0:07:46.13,Default,,0000,0000,0000,,because it's really hard when\Nthere's no contracting involved
Dialogue: 0,0:07:46.81,0:07:50.72,Default,,0000,0000,0000,,to guarantee regardless of\Nwhat the other party says then
Dialogue: 0,0:07:50.72,0:07:52.56,Default,,0000,0000,0000,,when it comes down to it\Nthey're actually going
Dialogue: 0,0:07:52.56,0:07:54.23,Default,,0000,0000,0000,,to cooperate given that it's
Dialogue: 0,0:07:54.23,0:07:57.29,Default,,0000,0000,0000,,in their interest\Nindividually to not cooperate.
Dialogue: 0,0:07:58.45,0:08:00.21,Default,,0000,0000,0000,,So one question that\Neconomists like to think
Dialogue: 0,0:08:00.21,0:08:04.06,Default,,0000,0000,0000,,about is then how can\Ncooperation be sustained
Dialogue: 0,0:08:04.06,0:08:04.98,Default,,0000,0000,0000,,in the real world?
Dialogue: 0,0:08:05.61,0:08:07.26,Default,,0000,0000,0000,,Well, one thing that's\Nimportant to remember here is
Dialogue: 0,0:08:07.26,0:08:11.54,Default,,0000,0000,0000,,that in the real world this\Ngame isn't played just once,
Dialogue: 0,0:08:12.03,0:08:13.50,Default,,0000,0000,0000,,when you have firms interacting
Dialogue: 0,0:08:13.50,0:08:16.96,Default,,0000,0000,0000,,with each other people making\Nthese decisions often times they
Dialogue: 0,0:08:17.02,0:08:19.68,Default,,0000,0000,0000,,have the chance to make the\Ndecisions over and over and over
Dialogue: 0,0:08:19.68,0:08:24.59,Default,,0000,0000,0000,,so when you have what's called\Na repeated game you might have a
Dialogue: 0,0:08:24.59,0:08:27.10,Default,,0000,0000,0000,,situation where people\Nstart testing out the waters
Dialogue: 0,0:08:27.25,0:08:30.20,Default,,0000,0000,0000,,to say well maybe if I\Ncooperate the other guy's going
Dialogue: 0,0:08:30.20,0:08:32.87,Default,,0000,0000,0000,,to cooperate and then\Nwe can keep this going
Dialogue: 0,0:08:33.93,0:08:39.11,Default,,0000,0000,0000,,because to cooperate here and\Nhope for the best outweighs,
Dialogue: 0,0:08:39.11,0:08:41.22,Default,,0000,0000,0000,,you know there's this\Nthreat of well if you try
Dialogue: 0,0:08:41.22,0:08:44.47,Default,,0000,0000,0000,,to screw me one time we're\Nreverting back here actually
Dialogue: 0,0:08:44.47,0:08:47.67,Default,,0000,0000,0000,,gives in the long term an\Nincentive to cooperate.
Dialogue: 0,0:08:48.84,0:08:51.52,Default,,0000,0000,0000,,So like I said it seems a little\Nbit artificial to be talking
Dialogue: 0,0:08:51.52,0:08:54.96,Default,,0000,0000,0000,,about this context of\Nprisoners being interrogated
Dialogue: 0,0:08:55.72,0:09:00.71,Default,,0000,0000,0000,,because really we're\Ntalking about economics.
Dialogue: 0,0:09:01.85,0:09:08.07,Default,,0000,0000,0000,,But it's very easy to see how\Nthis situation could be relevant
Dialogue: 0,0:09:08.07,0:09:10.76,Default,,0000,0000,0000,,in an economic context by just\Nreplacing the intuition behind
Dialogue: 0,0:09:10.76,0:09:11.28,Default,,0000,0000,0000,,some of the choices.
Dialogue: 0,0:09:11.28,0:09:14.01,Default,,0000,0000,0000,,So what I did here is\Nset up the identical game
Dialogue: 0,0:09:14.23,0:09:21.54,Default,,0000,0000,0000,,and have this model as\Nstill player 1 and player 2
Dialogue: 0,0:09:21.54,0:09:25.53,Default,,0000,0000,0000,,but now they have the\Nchoice of whether or not
Dialogue: 0,0:09:25.69,0:09:26.66,Default,,0000,0000,0000,,to cooperate or to compete.
Dialogue: 0,0:09:26.66,0:09:32.32,Default,,0000,0000,0000,,And you can see here they'd both\Ndo better off by cooperating
Dialogue: 0,0:09:32.32,0:09:35.16,Default,,0000,0000,0000,,but they also all have the\Nprivate incentive to compete.
Dialogue: 0,0:09:35.44,0:09:38.41,Default,,0000,0000,0000,,And you can notice here
Dialogue: 0,0:09:38.41,0:09:41.49,Default,,0000,0000,0000,,that this situation is\Nactually pretty realistic
Dialogue: 0,0:09:42.93,0:09:47.82,Default,,0000,0000,0000,,because at least in the United\NStates firms are not allowed
Dialogue: 0,0:09:47.85,0:09:49.23,Default,,0000,0000,0000,,to contract on whether or not\Nthey're going to cooperate,
Dialogue: 0,0:09:49.26,0:09:50.10,Default,,0000,0000,0000,,that's called collusion,\Nit's illegal.
Dialogue: 0,0:09:50.13,0:09:51.12,Default,,0000,0000,0000,,So they really are\Nsimultaneously making
Dialogue: 0,0:09:51.15,0:09:52.38,Default,,0000,0000,0000,,independent choices as\Nto how much to cooperate
Dialogue: 0,0:09:52.41,0:09:52.98,Default,,0000,0000,0000,,with their ''competitors''.