EEP100 - Lecture 28

Uploaded by calcommunitycontent on 11.12.2009

Is everybody ready to go? Yeah? Okay soÉ Once againÉcourse teaching and course evaluation
up there when you're done, and the course learning over here.
Thanks for surviving the course; I hope you enjoyed it so far. Today is a review day.
I'll get into some logistical stuff when I talk about extra credit. But first I'm going
to give you your bribes. And this is a lesson.
Anybody who has a peanut allergy; do not touch these things. They're cookies.
You should have brought those out before we did evaluations.
That's the lesson. So these are the peanut butter raisin things.
And there's two trays of those. And on top of them I'm going to put (in foil)
the walnut cranberry date things. And there are fewer of the walnut cranberry date.
We will see if there is actually an exchange and trading later going on. I will
let you guys decide the market price. Of course it's based on demand, which is based
on tastes. And there should be enough for one for everybody.
So take one, and if there's extra, then the gluttons can go to town. And try
and pry them out. I actually left them in the trays. I hope you like it. And pass these
around. And if you don't wantÉfeel free to not eat.
I ate them, and I'm still alive, so they're good.
I made them; what are you taking about! I ground the flour that went in these
things! My landlord's like, "You ground the flour? Woah."
It's Berkeley. So I did not give you your bribes before you
did the evaluations for an obvious reason. I don't want to influence you unduly,
but there is the notion that I could've passed it out and said, "No, this should not
influence you." That definitely is cheap talk. So I put my
cookies where my mouth is, and I excluded this from your contemplation in terms of my
course evaluation, which is not on my baking skills, but on my learning skills.
So hopefully that will be something good to eat, and the evals will be happy for everybody
who reads them. Is Jing here yet? Jing? Here's your regrade.
Somebody gave me a pen, but I don't know who the owner is. If they don't claim it
right now, I'm keeping it because it's kind of nice. ASUC? Anybody? No? Pen?
Okay, my property. So you've got your bribesÉextra credit, just
for clarification sake, everybody who's working on these projects. Number one: everything
has to be done before the final. Number two: if you do an extra credit project
of any variety, I want it printed out and handed in on the final.
There was one person that said, "I'm going to send you an audio recording file on e-
mail". That is acceptable. PowerPoint isÉsomeone had some PowerPointÉI don't
want to have 50 PowerPoint slides. So that's probably acceptable. But everything
else should be printed out and turned in before the final. If you're contacting
representativesÉjust as a clarification. You can contact (on the ethanol
question)Éyou can contact up to 2 representatives and 2 senators. You shall not
contact 52 of them and then write down the 2 that respond to you. So you put your
stick in the sand and you say, "I'm contacting these two." And that is your good or
your bad luck as far as them coming back to you. And you should put it on the sheet
yourself. Don't send me the names and say I should put it in the sheet. I understand
that you're sending the names as confirmation just in case some idiot decides to edit
away your property rights, but you should add your name to the spreadsheet. Any
questions onÉ So if they don't reply, you have toÉ
If they don't reply, it doesn't count. If they send you a form letter, it doesn't count.
If they say, "I love penguins." That's not a
form letter, but it still doesn't count. So I will
look at the replies that you get, and I will decide if you get between 1 and 5 points.
If Senator Diane Feinstein says, "That's a great
ideaÉI'm introducing legislation next week." then I will give you five points for
saving the world. And you'll be world famous moreover. Not just that. So intrinsic
and extrinsic motivation. Any other questions on the extra credit?
No? That's peanut butter. You guys can't talk with peanut butter in your mouths.
So they have to reply before the final? Yeah. There's a deadline, unfortunately. If
they reply the day after the final, and it's something awesome, send it to me. Let's say
it this wayÉI'm not guaranteeing any extra points for anything that happens after
5 pm on the whatÉthe 14th? 15th? 15th. Okay? 15th is when we're taking the final.
Everybody knows the final's at 50 Birge? Right? How did the public goods games go in discussion
last week? Somebody? Did it happen?
Yes. Cookies left. Okay, seconds!
Do I have to lead this discussion here? Okay, yes?
I have an interesting observation. SoÉmy group was really stingy during first round,
but during the second round, everybody was putting their money into the group bid, so
everyone would get their share of 200. But our group decided that we would rather
have someone in our group win over anyone else, so we were allowing who had the
most points from the previous round to keep their 50 and their individual ones to get
their share of the 150, and it was funny because I had been stingy the entire time first
round. First I was into it; I never put any points into the group pot, thinking everyone
else would do that? And then they still let me do that.
Okay, so what happened twoÉin phase one, and phase two.
Phase one is whereÉI want to go with your comment.
Phase one wasÉremind meÉeverybody was in a group, but you didn't know who
was in your group. It's anonymous And you had been free riding the whole time.
So you had the most points. Yes.
And then in the second phase, it was a known group. I can't remember. Was it
everybody scrambled up again? You were in new groups? Okay, that was good,
obviously. Otherwise people would've said, "No way, Jose."
No, you were in the same group as the first time.
You were? [student chatter]
Hold on. Say again? In my section, they were in the same group
as phase I. That was a mistake.
Ah. They should've been scrambled and started
again. It should've beenÉin phase I, you should've
been in an anonymous groupÉyou don't know who's there. In phase II, you should've
been sorted into new groups. Maybe I made a mistake in designing it.
It wouldn't really matterÉ It can create what's called a path dependency.
If everybody's been defecting in phase one, and they sayÉno you know everybody
who you've been working with, and they all defected last time, then what
happened in phase I would influence what happened in phase II.
But we weren't supposed to show what happened in phase I.
But everybody saw the group totals. It was on the opposite side of the page. So
no one knew what your individual contributions were.
But you know how your group did. Yeah.
Oh, okay. So let's say it this way. Maybe that was helpful. But then, on your group,
they basically said, whoever has the most points, we're going to let free ride, while
we contribute, so they can get the most points at the end of the session. So they're
basically saying, "We knew you were free riding from before, and now we're going
to reward you for it." Yeah.
Which is very laudable, I would say, in terms of sacrifice. On the other handÉit was
more likeÉat least somebody in our group is going to get the three points compared
to someone else. So in that sense, I guess it kind of does work in terms of travel
survival. So Fei nailed it. He was thinking exactly
the right thing. It's okay, you can do it either
way. Doesn't matter. I wasn't thinking that way.
But what would happenÉgenerally speaking; was there more free riding in phase I
or phase II? One
One. Mostly becauseÉnot the incentives would change, but the mechanisms of
negotiation had changed in a sense. You're looking at who's in your group, and
sayingÉeither you're talking to each other and it could be cheap talk, because you
could say, "Let's do this." And the decisions in phase II were still anonymous, as in
people in your group did not see what you did. Correct? Yeah?
But you could actually talk with either other and say, "can you please do this?"
And usually, people will say what they're going to do. Lying is much more difficult
when it's face to face. This is just important. It's a human thing. I put this on my
blog last week. Academic research that is so obvious thatÉwhy are we doing it?
Harvard Business School, surprisingly, or not, we're doing research showing that
when people negotiate face to face, they tend to be more cooperative. So then you
go from anonymous to known, and cooperation should go up. And generally
speaking, that the way generically what happened on average.
Till the last one. Until I the last one. That was even more interesting.
So everybody was cooperating all the way to the end, and bang it became
a one shot game. And apparently sometimes it was not even spoken that it became
a one shot game. Right? So that was interesting in terms of peoples' strategy.
Any other observations from that? It seemed like to be a winner, it was really
arbitrary, because it really depended on the other teammates. If they were more generous
then we would win, and if they weren't generous, then likeÉit wouldn't have worked
out anyway. Because likeÉI mean, for meÉI never contributed to the poolÉI always
kept my 50 points, and I'm sure the other winner did that as well, but she still ended
up with like 80 points more than me because her group was probably more generous.
Yes, but I'm not sureÉthat was in some ways a little bit arbitraryÉI thinkÉI don't
know if the word you use was arbitrary. But on the other hand it was not nearly as
arbitrary as the distribution of valuations on the auction game, right? Where some
people literally came out the gate with a high valuation. And it made it easier for
them to win. The mix of naughty versus nice or cooperators versus free riders that
you were working with is not something that we can control or hope to control,
right? But it could determine the outcomes. But it
was much more endogenous. There was much more endogeneityÉespecially in round
two. In these public goods games, as I mentioned
before, the economic prediction does not come true. The economic prediction is
that everybody free rides. If you get a group of economists, that's what happens.
But when you get a group of regular folks, or people that are in econÉlike you
guys, that understand the pay off and the trade off, then there tends to be a lot more
cooperation than zero. Right? And one of the best versions of this game, which has
been played hundreds, if not thousands, of times by academics, is when they took a
bunch of folksÉjust sayÉa group of everybody here, identified a group of freeriders,
put all the freeriders in a group with each other, or groups with each other,
and identified that you're in groups with freeriders, and then what happened?
They cooperated? They cooperated. They knew that theyÉtheir
strategy depended on having cooperators around them to make money, right?
But they all knewÉthey were all playing this strategy.
That's one thing that we know from these public goods games. Yes?
I wondered actuallyÉin phase IÉthere was sort of a humpÉif you got above that hump
as a group, you would go and put in more as a group, and the next time it would be
lowered. So I was in a group where I was the only one putting in something in the
beginning, and nobody else did, and I thoughtÉwell maybe the next time it will
actuallyÉand all the other groups already had moreÉ
Yes. But then actually my group, instead of going
up, got less and less. Because I guess everybody figured, well nobody's put stuff
in soÉ And that's because, in the end, a whole bunch
of people end up being reciprocators. They look at what other people do. Not just
in their group, but also what the otherÉ Did anybody look at what the other groups
were doing and change their behavior? Or just look? Did they change their behavior?
NoÉjust maybe? Let's say that it had some influence. Potentially,
not on the margin. It's maybe a curiosity.
I did some research on public goods games that wasÉa public goods game very
similar to what you guys played, and the computer screen...this was all computer.
It showed what you gave, what the group gave, and thenÉon one version of it, it
showed the average for not you. What everybody, but you, gave.
If you know these two pieces information, you can calculate this third thing here. I
actually did a version with only these two pieces of information, and how you do it?
You take the group, you subtract yourself, and you divide by the group number
minus one. Then you can find that average. On one version I showed the average, and on
one version I did not show the average. And the thing that was interesting on this
is that the massive switch in terms of the shares of freeriders, reciprocators, and cooperatorsÉ
When that information was shown, the number of reciprocators went up by a lot. It
went up by almost 20 points in terms of the share in the population. So the
reciprocators, I think, was something like 60% of the population, and it went to
80%. And, essentially, reciprocators are the people that look at this kind of
information, right? The information was implied in the first one,
but here it was explicit. And so when it was explicit, the number of reciprocators
went up. Now here's 2, I thought, pretty interesting footnotes. Number one is that
the average profits were identical. LikeÉwithin a point. And what that means,
in a sense, is that if you have these two groups, and they were actually competing,
to me, what it meansÉthis is calledÉsomething along the lines of evolutionary
stability. Evolutionary stable strategy. ESS. Anybody in ecology? Is that
what it is? Evolutionary stableÉanywayÉthese two groups were almost
identical in terms of their payoffs. And if you thought of that in terms of two
tribes that were competing with each other in terms of who is going to survive,
you could not pick a winner because their payoffs were almost identical. Does everybody
understand that? The payoffs as a group. Here's the thing that blew my mind.
That changed. That swap was almost entirely driven by women who were changing
their behavior. And if you wanted to be generous with your
explanation, you would say that women are the ones that balance out in different
situations to keep the tribe aliveÉin a sense that they are changing their behavior
to keep the payoffs. They didn't even know what the other groups
were doing. But they were much more responsive to the change in the dynamic in
the environment. I looked into the idea that girls can't do
math, which is another idea. That gets rejected, by the way. Literally, there's a
branch of experimental economics called gender. And so I wrote this paperÉthis paper
is a nightmare in terms of not getting published, but the result is very interesting,
and it was the girls that were switching their behavior and essentially keeping an
even keel for their tribe. So that's more information on how these public goods games
are played out. It's a very, very fertile area for research in terms of how we get along
as folks. Any other comments or questions on that discussion
section experience? Yes? It's just interesting because if you're playing
for something like 3 pointsÉI guess you want to just assume that everybody values
that the same way, or like wants the 3 points. Because I don't knowÉit's likeÉmy
strategyÉif you call it thatÉlike I don't knowÉit's just likeÉI know I'm not supposed
to put in 50 points each time to the group, so I guess I was probably really beneficial
to my group members, but in terms of returns, I guess I got a happy feeling
ofÉ Right, warm fuzzies.
But I guess the second time throughÉbut our group, just as a generalÉfor the first
round was like significantly higher than any of the other groups in terms of giving, and
then the second time around, I guess we did something similar to thatÉlike we just
gave it to whoever was going to winÉthat kind of thing? So I don't knowÉI guess it's
likeÉI just thought that was interesting. No, it is interesting. The whole point is
to think aboutÉyou're interacting with other people. This is the whole group dynamic. And
some people think that experiments are not necessarily useful. It's called external
validity. Can you do stuff in a lab and then go apply that outside. And they've actually
done these lab experiments with civil servants in Indonesia and students in
Indonesia on corruption. It turns out that the students were more corrupt
than the civil servants. They saidÉmaybe it's just cheap talkÉyou're just
doing it because it feels good. There's no real money at stake. They've played these
games when you have a week or a month of wages on the table. And people, in
especially things like the trust game, people will leave a month of wages on the
table rather than do the wrong thing. So these intrinsic motivations of people are
very strong, even when you put a wall of money as a motivation to defect.
In a sense, the behavior that you exhibit in these experiments tends to be behavior,
which is very similar to how you would work out in the real world. It's an ongoing
fight between experimental economists and what are called field experimenters.
But one way or the other, we're getting insight into how people interact and act, and
those experimental results are influencing the academicÉthe theoretical stuff that is
presented in econ 1. Or it should. A lot of econ textbooks don't reflect this stuff, and
they give you this kind of autistic homoeconomicus, who actually doesn't cooperate,
ever. That's one of the thingsÉthe points I've been making in this class.
Would that somewhat change if youÉlet's say this game is played with individuals who
each have feelings aboutÉfeelings that they want to be accepted or what if you play it
with big companies or corporations where the name of the companyÉnobody really
feels that's "me". And so likeÉsay you have four power companies each putting into
the public goodÉwould they act differently? Because there'sÉ
This is a point that a lot of people miss, I think, that companies are also people,
right? Now, you could haveÉwhat I mentioned last week is that when you have 2
people that are monitoring on a punishment game, that they will actually say it's
someone else'sÉit's his responsibilityÉoh, it's his responsibility. So you could have
that effect; you could recreate that effect very easily in the game with companies.
Companies being people. But if you put people in there and say you, individually,
represent Exxon, or whatever, then that person will take that personally, right? And
in the real world, of course, they aren't an individual, so you do see this kind
ofÉpassing the buck. Placing the blame on someone else. But that doesn't mean
that you can'tÉ You have to take that into account when you
go back into the real world. And if you want to put individual responsibility on big
people in corporations, then you say, "You are the face of this country." The way
that they did with Sarbanes-Oxley legislationÉwhere the CEO and the CFO would
have to sign the financial documents personally guaranteeing that they were valid,
or they would go to jail, personally. This is the kind of thing that does make a difference.
Another thing is that when all of the investment banksÉthey went from a
partnership structure to a corporate structureÉso Goldman Sachs went from partnerships
toÉwhich meansÉin a partnership structure, if a company goes broke,
you all lose your money. But corporate, they hadÉit's likeÉnow I'm
Goldman Sachs, and you're my shareholders, and don't worry, I'll take care
of you guys. But when Goldman or Lehman (more importantly)
goes bankrupt, then I keep my bonus, and you guys are broke, and that creates
a principle agent problem. Partnerships don't have that particular issue.
Other stuff on this? Right. Questions and answers on anything.
Sorry, I hadÉlike a fourth pointÉwith the gameÉ
Public goods game discussion. Go ahead. I've been reading articles aboutÉbecause
my research has been on likeÉIndia tribes and how they should like use their money,
and it took more time than I thought it would becauseÉI guessÉit'sÉone of the authors
wrote about how likeÉculture influences how people will give moneyÉthere's
like research done on culture and how that influences economic behavior at allÉlike
2 different ways of analyzing itÉbecause the thesis was that cultureÉlike within the
culture of the tribe legitimizes like the tribal council or whoever the government body
isÉ[inaudible] Well it depends onÉcultureÉeconomists have
avoided the word culture for a long time. Mostly, because you can't put it into
the math. But you knowÉwhen you have your utility function, and you could say it's
a function of the stuff you consume as reflected through this alpha parameter called
culture, thenÉyou mightÉ Alpha mightÉlet's just sayÉthis is what
you consume, this is "I", and then you could actually say, "What about everybody else consumes?"
and then let that reflect on a parameter called Alpha.
If you're a collective type of culture, then this "not I"Éthis weight on "not I" is
actually very strong. If you're in an individualistic culture, then your weight on "not
I" is weak. And this is well-understood. The problem is that in homoeconomicus, we
can ignore that, and that means we can do very simple math. But as soon as you bring
in the idea that you might actually start to worry about other people, then the
math falls apart. And so... "let's ignore that" has been the solution for a long time,
but that's wrong. You have to pay attention to peoples' attitudes towards other
people. It's not necessarily altruism either. There's a certain warm fuzzinessÉhow
I feel personally. I don't even care about that other person.
The way that people will donate money to international aidÉI think I mentioned this
beforeÉthey put the check in the envelope, and then they're happy. They don't care
if the check goes into the toilet, or if it's used for corrupt endsÉsex for food scandal
type of things. So that's the kind of stuff that really does matter when you're doing
research about how people interact. Yes? Good?
I had a questionÉI was reading Schelling andÉcan you clarifyÉ
You were reading Schelling? I was.
WowÉoh hold on. I want to do a survey question. Hold on to your question.
Who here read Frank? This is likeÉwe're all past evaluations, your courseÉyou can
be honest; I want to know. Who here liked Frank?
Was Frank the green one? The green one. The Economic Naturalist. The
green book! Who readÉoh! What was the second book called?
Was that Hazlitt? HazlittÉwho read Economics in One Lesson?
Okay, let's do it the other way around. Who didn't read the Economics in One Lesson?
Okay. That's okay; I just want to know.
Who read and liked it? Who read it and didn't like it? Okay, good.
And then Olson. Who read Olson? Who didn't read Olson?
Who read Olson and liked it? Who made it to the end? Who read Olson and didn't
like it? Okay. So you're stillÉyou had read it, of course.
Schelling. Who read Schelling? Reading it? Who intends to read Schelling? Who
intends to ace the final? Alright. Sorry. So Schelling?
I just wondering if you could clarify what he meant by conservative quantities as
opposed toÉ Oh, of course. What the hell. Let me see.
He's talking about ski liftsÉ Ski lifts? So the idea of having a closed
systemÉand he has a ski lift example in this. This is myÉI'm going to give you my brief
idea, but I could be wrong. So you have a ski lift, and the people are
going to go up. Let's say A are going to go up and B are going to come down. A will equal
B on a ski lift system. That is a closed loop.
If you have people helicoptering in, or people going off on some other trail, then A
will not equal B. So you have to worry about accounting for
where people go, and I'm not sure exactly what his point isÉit might be that there's
noise in the system? Because he mentioned that people can making
the lifts go faster would actually make everything to slowerÉ
It's like this crazyÉit's like it hurts your head to think about that, yeah. If you make
the lifts go faster, then everybody's going to show up here at once, but they all have
to come back down again. So you can never run faster than people are coming and
going down the hill. So in that senseÉspeeding up the ski liftsÉoh, that's rightÉso if
you're speeding up the ski lift, you still have to get everybody back up again before
they can get on the ski lift. And there might be a line here. If you slow down the ski
lift, there'll be a line. If you speed it up, there won't be anybody. So that's kind
of what I think is going on.
The book is essentially a discussion of dynamics though, right? There was a lot of
ramblingÉ He's a rambling guy. He's one of those non-math
guys who's not concise, but he discusses a lot of stuff, and he makes you
think about it. And this is the kind of stuff that you can't do with Schelling. You can't
mathematize Schelling, right? He was very famous for coming up with the concept
of "focal point" right? A focal point, for example, is thatÉif I
sell you guysÉwe're going to have the final somewhere on campus, not 50ÉwhateverÉand
not this classroom. And it's going to be at 5 o'clock on the 15th. But I'm not going
to tell you where it is, but it's going to be where we all are going to meet.
WhereÉif you had to meet somebodyÉ.this is the generic ideaÉif you had to meet
someone on campus, but you didn't know where you were going to meet them, but
you know when, and you knew that they knew that you you were trying to figure
out where they're going, and you're trying to figure out where they're going, because
they know that youÉthey're both from UC Berkeley. Where would you meet?
Where would you have the final exam? Now think about it for a second. And I want
5 hands up before IÉ Can I ask a question?
You can ask question, yea. So we can't meet here.
No. Not here, and not the official here. There's a hand, hold on.
I need five opinions I have a question. Can we communicate?
No, no you can't. You don't have cell phones anymore.
So, where? Sproul.
Sproul. Outside this building.
Outside this buildingÉnice try. Sproul
Sproul. Outside this building.
Sather Gate. Sather Gate.
I was thinking in the middle. What middle? What's the middle?
So this is actuallyÉthere is no focal point. There's a coordination problem. I was
actually thinking the Campinile. The tower. Me too.
Me too. See? You guys are right. A. But a focal pointÉand he gave the exampleÉyou've
got to meet somebody on New Years Day in New York City. Where would you
go? A focal point in that instance wasÉpeople
would say Time Square, which is a mess, or the clock at Grand Central Station. These
tend to be focal points. It's not 42nd and Broadway. I don't even know if those two cross.
So focal points are kind of a place where peopleÉyou thinkÉwhere would I go,
and where would you go, and if you know that I'm going thereÉyou see what I'm
saying? And
it can be very imperfect, given that we would have had a problem with
the final. But that's the idea that Schelling came up
with. NowÉthe reason I gave this example isÉnumber one: you should pay attention to
the idea of a focal point, but number two: you can't do math to figure out a focal
point. It's just this huge mess calledÉtrying to think of what everybody's
thinking. There's no math involved. And this is the thing Schelling specializes inÉthese
social dynamics, right? So, thinkingÉthe book is to have you think about
things. It might make your head hurt, but that's the idea.
It's just like the whole 50 people move here, and 80 people move thereÉlike the whole
back and forth thingÉ Yeah, it is a little bit of rambling because
it's justÉthe problem with the book is you can just keep typing away. ButÉsometimes
it does deserve a longer discussion. Sometimes a discussion is too long. Sometime's
it's too short. Other questions? Can I ask a question about thatÉrivalÉwhat
is the difference between a club and a pub? And why is oneÉ
Club good, public good. Oh, public.
Free as in beer. Any other questions? Do we have essay questions on the final?
No. The final exam is 16 true/false questions (none of which have been written by
me). And 7 or 8Éor 6 or 8 long questions that are similar to your homework
questions. Calculating, writing downÉmathematical type of stuff. Very similar to
the midterm. No blue books; we're going to give you the
paper. And you'll have plenty of time because it's three hours. This class session
is 80 minutes. That's 180 minutes, and it's only going to be 50% longer.
Other questions? What's in the top left box?
This is a private good. An apple, a can of Coke.
So SchellingÉis that behavioral economics? Or what would you classify that asÉ
Behavioral economics is essentially psychology relabeled as economics. And it has
to do withÉexperimental economics means doing experiments, which is using the
experimental methodÑthe scientific method. You can reproduce the experiment
over and over again. Behavioral economics has to do withÉit's
kind of strange just to say it's just like psychology, but you knowÉ
KahnemannÉI think that's the right spellingÉand Tueusky; these guys won the
Nobel Prize. WellÉI think Tueusky was dead. You can't win it when you're dead.
But Kahnemann supposedly did research on what was called "risk aversion". That is
something that I've mentioned to you guys before, right? There's risk neutrality,
there's risk aversion, and there's risk seeking. But what they found is that...when you're
looking at a win, you had, essentially, a typical utility function in terms of your
utility from a payoff of X. But then when you're looking at a loss, it
tended to be much more dramatic. The wholeÉin a senseÉ
Risk aversion is driven by loss aversion. So your utility fromÉthis is the loss here.
Minus x. And let's say it's an equal win, here. OrÉlet's be really dramatic. This
minus X here. And you have an equal win; so the utility of X compared to the
disutility of minus XÉthe disutility of minus X is much greater, right? If you have
this kind of shape of preferences, you are risk averse because you have a heavier
weight at disutility on losses. This is the kind of thing I would have a final
because this is the risk aversionÉwhat it means in terms of the definition.
Do we need to be able to identify likeÉwhich situations are complete information and
which one is incomplete information? Yeah, sure. That's a good question.
The incomplete one is when the player doesn't know the other person's action, or the
other person's payoff? Let's seeÉhelp me outÉ I don't even rememberÉcomplete
informationÉ.complete versus perfect. I thinkÉcompleteÉyou know
the payoffs. Is that right? Yeah. And then perfect, you know the moves. Is that
right? yeah. Do you know the payoffs too, in perfect information?
Not necessarily. You can have incomplete and perfect.
Would you know your own payoffs in perfect information?
No, you only know your moves. You may not know your payoffs. That's another 2x2.
For like...a poker game, would that be like incomplete because you knowÉlike the pot
that's in the middle, likeÉyou know how much you're going to give out? Or is it alsoÉ
You have to know the payoffs from a given action in order to know the complete
payoffs. Oh, but you're right. You don't know what the moves areÉno, but it's not
complete, because you don't know what hands people hold either. I'm not quite
sure what poker would be. It probably is incomplete and imperfect. It's probably
bad in that sense. Is prisoner's dilemma complete and perfect?
The prisoner's dilemma is complete and imperfect because you don't know what the
other person does as a move. In the example you gave with The Princess
Bride, the short man had complete information, but he didn't have perfect information?
Yeah, he didn't have perfect, and also didn't have complete. He thought he was
playing a game where there was one poison and one not.
He knows that if he doesn't drink the poison, he's going to get off with the girlÉand if
he does, then he's going to die. In that senseÉyeah, do you have an opinion?
He doesn't know the payoff of the other guy. Because for the other guy, he thinks that
if he drinks the poison, he diesÉif he doesn't, then he doesn't. ButÉ
That's true. He doesn't know the payoffs for the other guy either.
He thinks that if the other guy drinks the poison, he's also going to die.
He doesn't know that, right? He thanks that if he drinks the poison, he'll die. So he
thought there was one set of payoffs. That was not true. It was incomplete. And
even worse, he didn't knowÉhe thought he had perfect information because he
thought he only put the poison in one flask, and he put it in both. SoÉSicilians
shouldn't play these gamesÉincomplete and imperfect.
Anyone else? Can you go over the actual definition of risk
aversion? The actual definition of risk aversion is
thatÉ Or explain that over?
Explain this over? Okay. EssentiallyÉyes. Risk aversionÉI'll give you the math and
I'll talk about it. It means that the utility from a loss, or
the disutility from a loss, is greater than a
disutility from a win from the same amount, X. So, essentially, if you have a chance
of winning a dollarÉif I have a coin here, and I'm flipping a coin. And you have a
chance of winning a dollar or losing a dollarÉif you're risk averse, you're not going
to want to make that bet relative to not playing the game at all.
Not playing the game at all, you have a payoff of zero. If it's a 50/50 chance, the
expected payoff under the coin flip is zero. 50%, a dollar, 50%, not a dollar.
SoÉbecause your disutility from loss is greater than your utility from winning, the
expected win is zero, but your expected utility is less than zero. If you're risk
neutral, your expected utility is equal to zero because these are equal to each other.
That would mean that you would have this shape. That they're symmetric. You
could draw a straight line if you want to. But they would be symmetric. That's the
key. Does that answer your question? Other things? Other questions? Hold on.
Anybody else? You had 12. You had 12 too. Anybody else have a question?
I just wanted to sayÉabout the poker game? Yea.
I think that would be incomplete because there are a lot of odds when you play poker,
like if you really get into itÉlike everybody knowsÉ
I think poker is incomplete, imperfect. I thinkÉI meanÉ I'm ready to be persuaded
otherwise, but I think that you don't know the payoffs because you don't know
where you are in the game. You could calculate everything, but you don't know what
the hands are. Most professionals know all the odds.
Yeah. those are the payoffs, in a sense. Like expected payoffs.
That would make it complete. They know their own odds, but they don't knowÉwhat
they do is that they guess the cards of the other personÉ
That's the perfectÉthat's imperfect, but complete.
Wouldn't that be incomplete though? Because they're not sure what the other guy
hasÉ They know what the tree is; they just don't
know where they are on the tree. So this is complete. They know what the tree is, but
they don't know where they are on the tree. Complete, but imperfect. Let's just
take that as my opinion right now, but I think it's right. I'm willing to be persuaded
otherwise. Other hand, yeah? I think it would have to be incomplete for
this scenario. Why?
Because you don't know when somebody's going to go all-in. I meanÉyou could guess
where they're going to go. You're talking strictly payoffs? You can see what's in the
pot, butÉI meanÉit might be thatÉyou have no idea what the payoffs are going to be
worth. That's why you're always go in whenÉ No, but that's imperfect. You don't know their
moves. But in terms of payoffsÉyou mean how much
money are you going to winÉ.when really, how much money you're going to in
is in the pot, right? YeahÉI meanÉyou know what the moves that
are possible are. You just don't know where you are on that tree. So your uncertainty
in poker comes from not knowing what people are doing. If you sawÉeverybody's
cards are on the table. And you're paying poker, you would know exactly what
to do and what they were doing, right? So I think the problem is not knowing their
moves; which includes the cards they're given. Nature's playing in here too. Nature's
throwing the random stuff out there. So you don't know nature moves, either.
But if you see the cards on the table; there is no doubt about what's going on. So you
have complete information, but when your hands are in yourÉwhen you're holding
your cards, you have imperfect information also. Okay sorry. Another hand?
Sorry. Going back to this graphÉis it symmetrical for risk averse or risk neutral?
Risk neutral. It's asymmetric for risk averse, and with a heavier weight on losses. If
you're risk seeking, there's a heavier weight on wins, right? Okay, now. Our popular
two. You have a question still? It was just answered.
Awesome. For the risk aversion graph, which one is
it because it has two tails on the negative side?
No, no, no. I just drew another tail. A different tail. It's not two tails.
Why doesn't that concave up when it's negative? Wouldn't itÉbecause the rate of
decrease would be greaterÉ It's just meant to be symmetric. Just like
that. That's symmetric. If you put a greater weightÉif you have a
shape like thisÉthis is risk averse. And if
you have a shape like this, that's risk seeking. That's my point. Those are the three types.
So there is no inflection point? The inflection point is at zero, but not here.
This is risk neutral. Inflection point is when you change from one to the other.
For the seeking oneÉisn't it the same? NoÉit's the opposite of risk adverse.
No, no. I mean the concavity. I have no idea bout concavity. What?
Okay so the inflection point is when likeÉconcavity changes.
Right. But for seeking, it's the same from negative
to positive. No, this is meant to be like this. And if
this is a smooth curveÉoh wait, I see what you're saying. So you're saying it's actually
just smooth like that? Yeah.
Oh, do it against this point here. Think about it like that.
So this is actuallyÉnoÉthere is an inflection point going on here. I just drew it
badly. Draw it again. Utility. The expected payoff. So if you're
risk averseÉthis isÉ If you continue this line here, it would be
like that. So there's an inflection point there. This is risk seeking. Downside risk
is not as bad of a disutility as an upsideÉutility from win.
Other questions? What are the actual definitions?
The actual definitions. Like for risk seekingÉhow do you qualify
as risk seeking? You qualify as risk seekingÉlike I said with
the utility function there. Your utility of negative X (which is going
to be negative) plus the utility of positive X
is less than zero. That's risk averse. So if you take a risk, and then you lose,
and your satisfaction is greater than it would have gained from, that that's where youÉ
Yeah. That's what I said before about the coin flip. Your expected payoff from plus
a dollar minus a dollarÉyour expected X is
equal to zero. If you are risk averse, your expected utility of X is less than zero.
This would beÉif it was equal to zero it would be what?
Risk neutral Risk neutral.
And if it was greater than zero? Risk seeking. Awesome.
A different question besides poker and risks? Or not?
When you're talking about incomplete concepts, what were you saying about the
change in surplus? LikeÉdestroying surplus from the initial dealÉcan you modify the
contractÉ? I don't knowÉwhat was I saying? I can start
with incomplete contracts and we can see where I get to what you're talking about.
Is that the idea? Tell me about incomplete contracts. Okay.
SoÉ We're talking about stranded assets? That
kind of thing? Yeah. I know what incomplete contracts are,
but I don't know what the point was. Well the point of incomplete contractsÉif
you have contractÉyou've got point one, point two, etc. Because of information problems
(not action problems) you can never specify every action at every point
in time. Let alone that you might have a problem with knowing what happens. But let's
say you can't specify every action at every point in time.
SoÉyou and I make a dealÉI hire you to serve coffee at my restaurant. And you're
serving coffee, and you're getting paid a wage. That's the contract. Then the
restaurant catches on the fire. And I say, "Take care of the fire; I'm leaving." That's
not specified in the contract. There's an incomplete contract. You're going to be
like, "Wait, I'm going to do whatever I feel is right because I have noÉI have nothing
to affect me in my contract in terms of my action in conditions of fire."
So you default back toÉwhat would be in my best self-interest? That could mean: "I
love putting out fires" because I'm a fire fighter. Or it could mean: "Get the fuck out
of here". So that would be a problem with incomplete
contract. And we know, with contracts, we can never write complete contracts. I got
into one that was a little bitÉit was not an incomplete contract.
I was on eBay selling something, someone bought it; you win. And then they didn't
pay me. eBay saysÉgo through the adjudication process. Wait for five days. Be
really nice. Send it to our committee of not doing anything. I'm likeÉscrew that. So
I sent it off a second price offer to someone else who did not do anything. They
didn't accept the offer. And I have a friend who wants to buy it. And I send her an
e-mail. Do you want to buy it? Bang. Yes. Should have done that in the first place,
right? But now, eBay might come after me and say,
"You said you promised you would sell it to so and so." And now we're getting into
disputes, right? So who knows where that's going to go? Personally,
I'm just going to ignore it and move on. But that actually was a complete
contract that actually I just went away from, thinking that a contract is BS. I don't
want to deal with that. Because the contract defends eBay. It doesn't defend me.
But you have to figure out what to do when people go, essentially, out of equilibrium
behavior on what you would prefer. Now, the whole idea of a stranded asset (in
terms of a strategic game), is that you have the railroad track and it goes to the
mine, which is otherwise known as a big hole in the ground. And you have the first
step, which is that you have a mine and no track. And there's Mr. A and Mr. B. Mr.
A says, "This mine is not worth very much unless I start shipping away my ore."
And says to Mr. B, "I would like to make a deal with you; would you invest a huge amount
of fixed cost and then have an operating marginal cost, so that I will pay
you an average costÉ" soÉless than price (you get some profit) based on a certain specification
of quantity. So they make a deal. This, in fact, is a complete
contract as far asÉI build a railroad, I ship out your stuff, you pay me money. Now,
the problem is, in terms of justÉpure game theory, is that before the railroad is
built, B has lots of negotiating power called: "I will take your deal or not".
After the railroad is built, B has incurred the fixed cost. That's a sunk cost now. And
B has much less negotiating power. Does that make sense?
Yea. This is a hold up problem. The railroad's
built, and then Mr. A says, "Hey, you know what? I think we should renegotiate our contract.
I will pay you p less than p*." NowÉthere were some people in the class that
said, "Wait a second; now B is screwed in terms of the trackÉorÉsorryÉA
is screwed in terms of not having a mine that's running." But, as I mentioned, A can
just leave the ore in the mine. It was already there. But B has gone into debt, let's
say, to pay those fixed costs. So that's why B is more screwed than A in that ex-post
renegotiation. Okay.
Does that help? Yeah. I just had another thing about the surplus
then. Surplus? About what?
When you're talking about the painting example, if you want someone to paint the first
downstairs, you're saying thatÉ Paint? Oh yeah, painter, right, go ahead.
But you're saying that takes away some consumer surplus because they would've paid
more for that downstairs? Give me more about the example.
WellÉthat was it. Okay, kind of changing the contract in the
middle of it, and it takes away consumer surplus because of what?
Because you can charge more for the other roomÉin totalÉmore services?
I don't know if thatÉthrow out that example. Don't use that example, or whatever
you think it is, because I not sure what it is
Didn't you talk about the mechanic too? LikeÉyou bring the car to the mechanic, you
say fix the light, and then he goes in and he says, "Alright, you also need a new engine,
and I'll charge you this much"É That's not an incomplete contract problem.
That's an asymmetric information problem. Principle agent problem.
Okay. Yeah, that's principle agent.
So we can never have a complete contract. Basically, yes.
Do we need to know what was in the guest lecture by Ties?
Yes. What?
Do we need to know what was in the guest lecture by Ties. The Dutch guy. Oh shit.
Yeah, it's on video, don't worry. 30 hours of video, woo! I think you guys saw the
transcription note. It's kind ofÉyou can watch the lectures in Persian and Russian
and all kinds of stuff. The transcriptionÉnot that it's Brynna's problem, but the
technology is not exactly synchronized. The words kind ofÉthey show up early,
they show up late, whatever. But it is quite entertaining to watch the subtitles go by.
Any other questions? That'd be funny if you like swearÉ
I don't know what happens to the swear words; it'd be interesting, yeah. Did you
type in the swear words? Often, right? I said, hey, high fidelity. Any others?
Do you still have office hours this week? Oh, yeah. Office hours. Not today. That was
two questions. So my schedule is really horrible this week. Is Diana still here? Are
youÉyou're not having any office hoursÉyou're clear, right?
Right. Between now and the final?
Yeah. So Diana's not doing anything. Fei is doing
what? I'm doing my regular office hours Wednesday
and Friday, and extra office hours on Monday.
Okay, so you guys know that? So extra office hours MondayÉwhen?
2-4. I sent an e-mail to everyone. You sent an e-mail; that's right. Just in
case people aren't paying attention. The day before the final.
I haveÉI can doÉeither office hoursÉI can do office hours and a review session, or
just office hours, or just a review session. And I don'tÉthe timing is of interest. So
you're doing this thing on Monday, right? Right.
So I think a review session closer to the final is often better. Do people want to do
a review session? Number one: a review sessionÑyes
or no? Yes? Okay, that would be yes. And on Monday or on Tuesday? Monday?
Tuesday? Last minute people. Okay. the final. The day of the final. No,
on the final. During the final.
Okay, it's Monday. Now Fei is doing something from 2 to 4.
What about Sunday? SundayÉI just worked Sunday. I'm not working
two Sundays in a row. Although I am workingÑjust not on your stuff. So Monday
at 4pm? A review sessionÉI'm assuming one hour or two hours?
Two. Any surprise there? So we've got a choice.
We've got 4 pm, which is after Fei's, or we haveÉlet's just say 10am.
10am. 10am say yay? 4pm say yay? Whoa. Come on guys.
Oh, I have to get a room, though. I have to get a room. This is a problem
because I can't get this room necessarily.
Just use someone's apartment. It'll be a BYOB review session. I don't even
want to know about your licenses. So are there any more suggestions on the floor
about a time better than these two? 12?
12. I'm swimming at 12. I'm busy. Who thinks 12? Okay, see? Forget about it.
Everybody's swimming at 12. Hard core 10 o'clock? If you're serious. You
only get to vote for one. 4 o'clock? Okay, 4 o'clock. Who can't make 4 o'clock?
Why? I have a final.
You have a final. Work.
WorkÉwhatever. I have another review session.
Another review session. Should we move it to 5? 5-7? Is that better?
There's a final at 5. There's a final at 5. So 4 is better. Okay.
This is last minute. Called: "I was reviewing it, and I have no idea about this." if you
start studying after 4pm then that's probably okayÉI don't know. Should be fine. I
know the material. Is this the same stuff Diana did yesterday,
or is this likeÉ This is whateverÉI'm going to show up, and
you guys are going to ask me questions. I am not leading anything. I'm going to show
up and talk shit. Or not. ShitÉbut it's relevant shit. On the final. You can ask,
"Oh, is this on the final." And I'll say, "I don't
know because I've already written the questionsÉ" So office hoursÉI can do office hours on
Thursday morning, this week, atÉanybody? 10 o'clock or 11 o'clock? Good or no? Does
anybody care? If you care, raise your hand. 11 o'clock? 10 o'clock? 10 o'clock on
Thursday, I'll have office hours. Any other questions?
Are we getting the evaluations back, ever? The briefings will come back at the final.
Okay. Your homeworks are back, your grades will
be uploaded, etc. Okay, great. See you guysÉeither on Thursday
or on Monday or on Tuesday. [Applause]
Thank you very much. Get a cookie, damn it.