1
00:00:00,090 --> 00:00:02,430
The following content is
provided under a Creative

2
00:00:02,430 --> 00:00:03,850
Commons license.

3
00:00:03,850 --> 00:00:06,060
Your support will help
MIT OpenCourseWare

4
00:00:06,060 --> 00:00:10,150
continue to offer high quality
educational resources for free.

5
00:00:10,150 --> 00:00:12,690
To make a donation or to
view additional materials

6
00:00:12,690 --> 00:00:16,620
from hundreds of MIT courses,
visit MIT OpenCourseWare

7
00:00:16,620 --> 00:00:17,860
at ocw.mit.edu.

8
00:00:27,546 --> 00:00:29,670
ANDREW LO: Actually, at
the end of class last time,

9
00:00:29,670 --> 00:00:34,020
we were talking about what
happened with the stock

10
00:00:34,020 --> 00:00:37,440
prices of those
four companies that

11
00:00:37,440 --> 00:00:41,730
were engaged in manufacturing
parts for the space shuttle.

12
00:00:41,730 --> 00:00:44,370
And I didn't realize
this, but apparently,

13
00:00:44,370 --> 00:00:46,800
in some of your
other classes, I've

14
00:00:46,800 --> 00:00:50,580
been told that you actually
did a case on Morton Thiokol,

15
00:00:50,580 --> 00:00:53,250
and the fact that there were
some interesting discussions

16
00:00:53,250 --> 00:00:55,350
between the Morton
Thiokol and NASA

17
00:00:55,350 --> 00:00:57,540
engineers about
whether to launch.

18
00:00:57,540 --> 00:01:00,840
And that, in fact, when
you just look at it

19
00:01:00,840 --> 00:01:05,180
from the perspective
of the data,

20
00:01:05,180 --> 00:01:08,640
it would not have made sense
to go ahead with the launch.

21
00:01:08,640 --> 00:01:11,220
Well, the point about
efficient markets

22
00:01:11,220 --> 00:01:15,600
is that, while it may not be
immediately apparent as to what

23
00:01:15,600 --> 00:01:19,800
all of the inputs are for
coming up with a bottom line

24
00:01:19,800 --> 00:01:24,360
price for an asset, the fact is
that enormous, enormous amounts

25
00:01:24,360 --> 00:01:29,100
of information are collected,
collated, analyzed,

26
00:01:29,100 --> 00:01:32,280
and ultimately
reflected in the price,

27
00:01:32,280 --> 00:01:35,940
to the point where, after
a six month investigation,

28
00:01:35,940 --> 00:01:38,940
it was determined that
Morton Thiokol was

29
00:01:38,940 --> 00:01:42,300
at fault for the explosion.

30
00:01:42,300 --> 00:01:46,950
But, in fact, within hours after
the crash and literally minutes

31
00:01:46,950 --> 00:01:49,230
we see that the
Morton Thiokol stock

32
00:01:49,230 --> 00:01:51,420
price was the one hit hardest.

33
00:01:51,420 --> 00:01:54,450
And even when you control
for other factors like size

34
00:01:54,450 --> 00:01:58,560
and industry effects, you
still have a pretty significant

35
00:01:58,560 --> 00:02:04,530
finger pointed at Morton Thiokol
purely from the stock market.

36
00:02:04,530 --> 00:02:07,930
In a nutshell, this is what
efficient markets is about.

37
00:02:07,930 --> 00:02:10,780
And for the longest
time that is really

38
00:02:10,780 --> 00:02:14,560
what academic finance
taught, and what

39
00:02:14,560 --> 00:02:17,580
was the basis for virtually
everything that we do.

40
00:02:17,580 --> 00:02:19,990
So, in fact, all the
material that we've

41
00:02:19,990 --> 00:02:24,520
covered in this course
really assumes and requires

42
00:02:24,520 --> 00:02:27,130
that markets work well.

43
00:02:27,130 --> 00:02:31,240
That markets reflect all
available information.

44
00:02:31,240 --> 00:02:34,720
Because obviously if it
didn't, then when I tell you,

45
00:02:34,720 --> 00:02:37,260
go look at the market, I'm
telling you to waste your time,

46
00:02:37,260 --> 00:02:37,760
right?

47
00:02:37,760 --> 00:02:40,150
Because why would
you consult something

48
00:02:40,150 --> 00:02:44,290
that really didn't contain
some significant informational

49
00:02:44,290 --> 00:02:46,250
advantages?

50
00:02:46,250 --> 00:02:47,660
That's the good news.

51
00:02:47,660 --> 00:02:50,360
That's the positive side
of efficient markets.

52
00:02:50,360 --> 00:02:53,000
Over the last 15,
20 years, there

53
00:02:53,000 --> 00:02:55,040
have been more
and more criticism

54
00:02:55,040 --> 00:02:58,510
leveled against this hypothesis
of efficient markets.

55
00:02:58,510 --> 00:03:03,410
And the most easy, the most
obvious and easiest way

56
00:03:03,410 --> 00:03:06,350
to make that point,
is with a joke

57
00:03:06,350 --> 00:03:09,950
that you might have
heard about economists.

58
00:03:09,950 --> 00:03:12,050
Two economists walking
down the street.

59
00:03:12,050 --> 00:03:15,620
One of them sees a $100
bill lying on the ground

60
00:03:15,620 --> 00:03:18,800
and starts to pick it up when
the other economist says, "Oh,

61
00:03:18,800 --> 00:03:20,720
don't bother picking it up."

62
00:03:20,720 --> 00:03:22,945
And then the first economist
says, "Well, why not?"

63
00:03:22,945 --> 00:03:24,320
The second economist
says, "Well,

64
00:03:24,320 --> 00:03:27,350
if that were a real,
genuine $100 bill,

65
00:03:27,350 --> 00:03:28,850
someone would have
already taken it,

66
00:03:28,850 --> 00:03:31,670
so it must be a counterfeit."

67
00:03:31,670 --> 00:03:32,880
Now you laugh at that.

68
00:03:32,880 --> 00:03:35,300
But that's exactly what
efficient market says.

69
00:03:35,300 --> 00:03:38,930
It says that you can't
make any money in markets

70
00:03:38,930 --> 00:03:42,650
because, if you could, someone
would have already done it.

71
00:03:42,650 --> 00:03:45,860
Markets are so competitive
that, in fact, you

72
00:03:45,860 --> 00:03:48,110
can't make any money.

73
00:03:48,110 --> 00:03:52,160
And after awhile, people
started asking questions

74
00:03:52,160 --> 00:03:53,450
about this hypothesis.

75
00:03:53,450 --> 00:03:56,550
In particular, they would
ask the question, well,

76
00:03:56,550 --> 00:04:00,140
if nobody can make
any money, then why

77
00:04:00,140 --> 00:04:03,590
does anybody bother going
into financial markets?

78
00:04:03,590 --> 00:04:06,080
If nobody makes any
money, then who's

79
00:04:06,080 --> 00:04:09,200
the folks out there that
is gathering information

80
00:04:09,200 --> 00:04:11,120
to make markets efficient?

81
00:04:11,120 --> 00:04:13,700
So first of all, there's
a bit of a Zen paradox.

82
00:04:13,700 --> 00:04:17,300
And we'll talk about that Zen
paradox a little bit later on.

83
00:04:17,300 --> 00:04:19,610
But the more
substantial criticisms

84
00:04:19,610 --> 00:04:24,800
have to do with the fact that
this presumption of efficiency

85
00:04:24,800 --> 00:04:28,670
requires rationality.

86
00:04:28,670 --> 00:04:31,970
In particular, it requires
that market participants

87
00:04:31,970 --> 00:04:34,910
behave the way that economists
would argue that they behave.

88
00:04:34,910 --> 00:04:39,770
So all of you are very
gracious in not arguing with me

89
00:04:39,770 --> 00:04:44,320
that you would all want to
hold the tangency portfolio.

90
00:04:44,320 --> 00:04:48,100
Now, you could have asked
me at that time, well,

91
00:04:48,100 --> 00:04:50,320
why would I want to hold
the tangency portfolio?

92
00:04:50,320 --> 00:04:55,270
I may happen to enjoy holding
a small subset of stocks

93
00:04:55,270 --> 00:04:59,260
that I know, and love, and
have a particular emotional

94
00:04:59,260 --> 00:05:00,940
attachment to.

95
00:05:00,940 --> 00:05:03,340
But you probably recognize
that that wasn't where

96
00:05:03,340 --> 00:05:06,730
the class was going, so you were
very gracious in not raising

97
00:05:06,730 --> 00:05:09,800
that as an issue, even though,
in your heart of hearts,

98
00:05:09,800 --> 00:05:11,290
you might really prefer that.

99
00:05:11,290 --> 00:05:15,040
You might simply want to hold
a few stocks that you know,

100
00:05:15,040 --> 00:05:18,710
and enjoy, and believe in,
and so on and so forth.

101
00:05:18,710 --> 00:05:23,000
Well, that's an example of an
irrationality from the point

102
00:05:23,000 --> 00:05:25,520
of view of the economist.

103
00:05:25,520 --> 00:05:27,500
And in a minute, I'm
going to actually call

104
00:05:27,500 --> 00:05:32,420
into question what we mean
by rationality altogether.

105
00:05:32,420 --> 00:05:34,790
So by the end of
this lecture, I'm

106
00:05:34,790 --> 00:05:37,064
going to make a fairly
strong prediction.

107
00:05:37,064 --> 00:05:38,480
I'm going to predict
that what I'm

108
00:05:38,480 --> 00:05:41,240
going to tell you in the
next hour and 15 minutes

109
00:05:41,240 --> 00:05:44,720
will change your
lives permanently.

110
00:05:44,720 --> 00:05:46,690
That's a tall order.

111
00:05:46,690 --> 00:05:48,470
But you'll let me know.

112
00:05:48,470 --> 00:05:51,230
In five or 10 years, I want you
to write me back and tell me

113
00:05:51,230 --> 00:05:52,370
whether or not I'm right.

114
00:05:52,370 --> 00:05:54,770
I think I am right,
because I'm going

115
00:05:54,770 --> 00:05:59,912
to change the way all of you
look at rationality completely.

116
00:05:59,912 --> 00:06:01,370
And actually we're
going to do this

117
00:06:01,370 --> 00:06:03,140
in the context of
your own behavior.

118
00:06:03,140 --> 00:06:05,020
All right?

119
00:06:05,020 --> 00:06:07,460
So let me tell you where
we're going with this.

120
00:06:07,460 --> 00:06:09,830
Behavioral finance,
the alternative

121
00:06:09,830 --> 00:06:14,540
to market efficiency, says that
market participants are simply

122
00:06:14,540 --> 00:06:15,190
irrational.

123
00:06:15,190 --> 00:06:16,580
They don't make
decisions the way

124
00:06:16,580 --> 00:06:19,460
that economists would
presuppose that they do.

125
00:06:19,460 --> 00:06:24,320
In particular, they suffer
from all sorts of biases,

126
00:06:24,320 --> 00:06:25,760
and they're listed here.

127
00:06:25,760 --> 00:06:27,470
Loss aversion,
anchoring, framing.

128
00:06:27,470 --> 00:06:28,250
Overconfidence.

129
00:06:28,250 --> 00:06:29,420
Overreaction.

130
00:06:29,420 --> 00:06:30,260
Herding.

131
00:06:30,260 --> 00:06:31,820
Mental accounting.

132
00:06:31,820 --> 00:06:32,534
And so on.

133
00:06:32,534 --> 00:06:33,950
And some of you,
I think, may have

134
00:06:33,950 --> 00:06:36,200
come across these in
some of the other courses

135
00:06:36,200 --> 00:06:37,880
that you've taken.

136
00:06:37,880 --> 00:06:40,670
If you ever take an
experimental psychology course,

137
00:06:40,670 --> 00:06:43,760
you would actually see
these kinds of examples

138
00:06:43,760 --> 00:06:47,600
in pretty intimate detail.

139
00:06:47,600 --> 00:06:52,430
When you look down this list,
and you take seriously what

140
00:06:52,430 --> 00:06:54,440
the psychologists
are telling us,

141
00:06:54,440 --> 00:06:56,885
you would conclude
that the human animal

142
00:06:56,885 --> 00:06:59,780
is among the stupidest
creatures on earth

143
00:06:59,780 --> 00:07:03,530
because we engage in all
of these kinds of biases.

144
00:07:03,530 --> 00:07:05,510
And I want to go
through a few examples.

145
00:07:05,510 --> 00:07:07,310
Some of you may have
already seen this

146
00:07:07,310 --> 00:07:09,290
because I actually
gave these examples

147
00:07:09,290 --> 00:07:10,859
in my introductory
talk on finance.

148
00:07:10,859 --> 00:07:13,400
And if you have, we're going to
go through it pretty quickly.

149
00:07:13,400 --> 00:07:15,691
But if you haven't, I think
you'll be kind of surprised

150
00:07:15,691 --> 00:07:17,180
by some of the outcomes.

151
00:07:17,180 --> 00:07:18,860
So the loss aversion
example is one

152
00:07:18,860 --> 00:07:21,230
that I did give in my
introductory lecture.

153
00:07:21,230 --> 00:07:24,320
How many people have seen
this or remember this?

154
00:07:24,320 --> 00:07:25,020
OK.

155
00:07:25,020 --> 00:07:26,978
So I'm going to go through
this pretty quickly.

156
00:07:26,978 --> 00:07:29,990
This is an example where clearly
most people, when they're

157
00:07:29,990 --> 00:07:36,560
confronted with a choice of a
certain gain versus a riskier

158
00:07:36,560 --> 00:07:39,920
alternative that might
have even bigger gains

159
00:07:39,920 --> 00:07:44,390
but might also entail
gaining nothing,

160
00:07:44,390 --> 00:07:48,800
that most everybody would pick
choice A over B, as most of you

161
00:07:48,800 --> 00:07:52,440
did when we ran this example at
the beginning of the semester.

162
00:07:52,440 --> 00:07:56,570
So even though B has a higher
expected value, more risk,

163
00:07:56,570 --> 00:08:01,550
more expected return, like
the CAPM says you should have,

164
00:08:01,550 --> 00:08:05,060
nevertheless, most of you
as a personal preference

165
00:08:05,060 --> 00:08:08,610
would pick the less
risky alternative, A.

166
00:08:08,610 --> 00:08:10,550
Now, when you're
confronted with losses,

167
00:08:10,550 --> 00:08:14,090
as you recall from the
example, most of you

168
00:08:14,090 --> 00:08:17,410
picked not the sure
loss but rather

169
00:08:17,410 --> 00:08:22,850
the gamble that actually might
cause you to lose even more

170
00:08:22,850 --> 00:08:25,550
but might also allow
you to get back to even.

171
00:08:25,550 --> 00:08:29,390
So the reason that these two
choices are so interesting

172
00:08:29,390 --> 00:08:32,690
is that, in the first
case of A versus B,

173
00:08:32,690 --> 00:08:35,690
you picked the less risky
alternative because you are

174
00:08:35,690 --> 00:08:38,539
being confronted with gains.

175
00:08:38,539 --> 00:08:41,570
But in this case, when you're
being confronted with two

176
00:08:41,570 --> 00:08:46,400
losses, instead of being
consistent and picking the less

177
00:08:46,400 --> 00:08:51,230
risky alternative-- which, by
the way, is the sure loss C--

178
00:08:51,230 --> 00:08:54,620
virtually all of you
picked the gamble.

179
00:08:54,620 --> 00:08:57,080
You're willing to
take more risk when

180
00:08:57,080 --> 00:09:02,450
it comes to losses than you
are when it comes to gains.

181
00:09:02,450 --> 00:09:04,730
That's an asymmetry
that frankly doesn't

182
00:09:04,730 --> 00:09:07,850
jibe with modern
financial theory,

183
00:09:07,850 --> 00:09:12,110
with utility maximization,
with efficient markets.

184
00:09:12,110 --> 00:09:15,260
So the reason that these
two sets of choices

185
00:09:15,260 --> 00:09:18,320
are so interesting when
juxtaposed together

186
00:09:18,320 --> 00:09:23,030
is that the choices that most
of you did pick, A and D,

187
00:09:23,030 --> 00:09:27,050
was actually strictly dominated
by the choices that most of you

188
00:09:27,050 --> 00:09:31,730
did not pick, which is C and
B. When you combine these two,

189
00:09:31,730 --> 00:09:34,580
you can show that B
and C is exactly equal

190
00:09:34,580 --> 00:09:38,600
to the most popular choices,
A and D, plus $10,000 cash.

191
00:09:38,600 --> 00:09:41,950
So you guys left $10,000
lying on the sidewalk.

192
00:09:41,950 --> 00:09:45,620
And that's clearly
not efficient.

193
00:09:45,620 --> 00:09:51,590
Well, it turns out that there is
a reason for this kind of bias,

194
00:09:51,590 --> 00:09:55,770
and it's hardwired
into all of you.

195
00:09:55,770 --> 00:09:58,310
And I'm going to tell you
what that is in a few minutes,

196
00:09:58,310 --> 00:10:00,500
after I go through just a
couple of other examples

197
00:10:00,500 --> 00:10:03,170
to make sure that
we all understand

198
00:10:03,170 --> 00:10:07,490
the limitations of what we
consider to be rational,

199
00:10:07,490 --> 00:10:10,020
economic decision-making.

200
00:10:10,020 --> 00:10:11,860
Any questions
about this example?

201
00:10:11,860 --> 00:10:13,230
This is called loss aversion.

202
00:10:13,230 --> 00:10:14,850
This is an example
that was first

203
00:10:14,850 --> 00:10:17,910
proposed by two psychologists,
Kahneman and Tversky,

204
00:10:17,910 --> 00:10:21,390
and for which Kahneman won
the Nobel Prize in economics

205
00:10:21,390 --> 00:10:23,532
a couple of years ago.

206
00:10:23,532 --> 00:10:25,740
Tversky would have won it
as well, had he been alive,

207
00:10:25,740 --> 00:10:27,700
but he died a few years before.

208
00:10:27,700 --> 00:10:29,400
And what's interesting
about the fact

209
00:10:29,400 --> 00:10:31,020
that Kahneman won
the Nobel Prize is

210
00:10:31,020 --> 00:10:34,590
that he's not an economist,
he's a psychologist.

211
00:10:34,590 --> 00:10:38,460
And so, in a way, maybe
the Nobel Memorial Prize

212
00:10:38,460 --> 00:10:42,049
in Economics was trying
to signal to the economics

213
00:10:42,049 --> 00:10:44,340
profession that they ought
to take these kind of things

214
00:10:44,340 --> 00:10:46,080
more seriously.

215
00:10:46,080 --> 00:10:46,830
OK.

216
00:10:46,830 --> 00:10:48,270
This is the second
example that I

217
00:10:48,270 --> 00:10:50,340
gave in that
introductory lecture

218
00:10:50,340 --> 00:10:52,710
on drawing balls out of an urn.

219
00:10:52,710 --> 00:10:55,720
How many people still
remember this example?

220
00:10:55,720 --> 00:10:56,490
No?

221
00:10:56,490 --> 00:10:59,040
OK, so let me go through this
in a little bit more detail

222
00:10:59,040 --> 00:11:00,930
because this is
directly relevant

223
00:11:00,930 --> 00:11:04,184
to the current financial
crisis that we're facing.

224
00:11:04,184 --> 00:11:06,600
Those of you who've ever taken
an undergraduate statistics

225
00:11:06,600 --> 00:11:09,025
course, you'll remember
the bane of your existence

226
00:11:09,025 --> 00:11:09,900
during that semester.

227
00:11:09,900 --> 00:11:12,340
It was the prototypical
urn example, right?

228
00:11:12,340 --> 00:11:13,980
Drawing balls out of an urn.

229
00:11:13,980 --> 00:11:15,250
I hated those examples.

230
00:11:15,250 --> 00:11:16,836
They're a real pain in the neck.

231
00:11:16,836 --> 00:11:18,210
So I'm going to
give you one now.

232
00:11:18,210 --> 00:11:21,270
But this one is not
as bad as the others

233
00:11:21,270 --> 00:11:23,070
that you might have come across.

234
00:11:23,070 --> 00:11:25,620
I want you to imagine
an urn that I have

235
00:11:25,620 --> 00:11:27,720
where I've got 100 balls.

236
00:11:27,720 --> 00:11:29,979
And 50 of those balls
are colored red,

237
00:11:29,979 --> 00:11:31,270
the other 50 are colored black.

238
00:11:31,270 --> 00:11:33,830
They're otherwise identical.

239
00:11:33,830 --> 00:11:36,780
And you and I, we're
going to play a game.

240
00:11:36,780 --> 00:11:38,180
And here's the game.

241
00:11:38,180 --> 00:11:39,749
You pick a color, red or black.

242
00:11:39,749 --> 00:11:40,790
Don't tell me what it is.

243
00:11:40,790 --> 00:11:42,680
Write it down on
a piece of paper.

244
00:11:42,680 --> 00:11:46,130
And then I'm going to draw
a ball out of this urn.

245
00:11:46,130 --> 00:11:49,740
And if the ball that
I draw is your color,

246
00:11:49,740 --> 00:11:52,430
I'm going to pay you $10,000.

247
00:11:52,430 --> 00:11:56,413
And if it's not, I'm
going to pay you nothing.

248
00:11:56,413 --> 00:11:57,740
OK?

249
00:11:57,740 --> 00:11:59,360
That's the game.

250
00:11:59,360 --> 00:12:02,600
And we're going to
do this exactly once.

251
00:12:02,600 --> 00:12:04,400
Play this game once.

252
00:12:04,400 --> 00:12:06,900
I wish we could do
this in reality.

253
00:12:06,900 --> 00:12:09,050
But the prize is a
little bit higher

254
00:12:09,050 --> 00:12:13,100
than what our dean is willing
to let me gamble with.

255
00:12:13,100 --> 00:12:16,640
But let's imagine we're
going to do this for real.

256
00:12:16,640 --> 00:12:17,540
OK?

257
00:12:17,540 --> 00:12:21,210
So I want you to tell
me, first of all,

258
00:12:21,210 --> 00:12:24,290
how many people would be willing
to play this game with me

259
00:12:24,290 --> 00:12:25,790
for $1?

260
00:12:25,790 --> 00:12:28,080
In other words, you pay me
$1 and we'll play this--

261
00:12:28,080 --> 00:12:28,620
Yeah, OK.

262
00:12:28,620 --> 00:12:29,190
Good.

263
00:12:29,190 --> 00:12:31,620
How about $1,000?

264
00:12:31,620 --> 00:12:34,290
$2,000?

265
00:12:34,290 --> 00:12:35,820
$3,000?

266
00:12:35,820 --> 00:12:37,740
$4,000?

267
00:12:37,740 --> 00:12:40,200
$4,500?

268
00:12:40,200 --> 00:12:43,320
$4,999?

269
00:12:43,320 --> 00:12:45,900
$5,010?

270
00:12:45,900 --> 00:12:46,800
OK.

271
00:12:46,800 --> 00:12:49,410
So I still had a few
hands up at $4,999.

272
00:12:49,410 --> 00:12:53,010
Nobody wants to play
this with me for $5,010,

273
00:12:53,010 --> 00:12:54,510
presumably because?

274
00:12:54,510 --> 00:12:55,353
Louis?

275
00:12:55,353 --> 00:12:57,492
AUDIENCE: So if you
asked the question,

276
00:12:57,492 --> 00:12:59,700
how many people would sit
down at a blackjack table--

277
00:12:59,700 --> 00:13:00,066
ANDREW LO: Yeah.

278
00:13:00,066 --> 00:13:01,357
AUDIENCE: Or a roulette wheel--

279
00:13:01,357 --> 00:13:02,174
ANDREW LO: Yeah.

280
00:13:02,174 --> 00:13:03,550
AUDIENCE: I think people
would raise their hand.

281
00:13:03,550 --> 00:13:04,292
ANDREW LO: Mm-hmm.

282
00:13:04,292 --> 00:13:05,422
AUDIENCE: But they
won't pay $5,010

283
00:13:05,422 --> 00:13:07,010
to play this game of red black?

284
00:13:07,010 --> 00:13:08,390
ANDREW LO: Well, that's a
different question though,

285
00:13:08,390 --> 00:13:09,160
right?

286
00:13:09,160 --> 00:13:10,040
AUDIENCE: This is
a roulette wheel.

287
00:13:10,040 --> 00:13:10,950
ANDREW LO: Well, but it's--

288
00:13:10,950 --> 00:13:11,810
AUDIENCE: [INAUDIBLE]
house money.

289
00:13:11,810 --> 00:13:12,980
ANDREW LO: --not a
roulette wheel though.

290
00:13:12,980 --> 00:13:14,330
It's a different game.

291
00:13:14,330 --> 00:13:17,100
It's an urn with 50
balls red and 50 black.

292
00:13:17,100 --> 00:13:19,022
AUDIENCE: [LAUGHTER]

293
00:13:19,022 --> 00:13:20,480
ANDREW LO: I mean,
you know, I have

294
00:13:20,480 --> 00:13:22,850
to tell you that a
roulette wheel happens

295
00:13:22,850 --> 00:13:23,869
to be more entertaining.

296
00:13:23,869 --> 00:13:25,910
You know, you see the
little ball rolling around,

297
00:13:25,910 --> 00:13:28,495
and it's a different game.

298
00:13:28,495 --> 00:13:30,620
All right, I want to come
back to this in a minute.

299
00:13:30,620 --> 00:13:32,600
But let's get back to the price.

300
00:13:32,600 --> 00:13:35,330
I established the
price of $5,000.

301
00:13:35,330 --> 00:13:37,520
And, not surprisingly,
all of you

302
00:13:37,520 --> 00:13:39,530
extremely numerate
MIT students, you

303
00:13:39,530 --> 00:13:42,890
figured out that the expected
value is $5,000, right?

304
00:13:42,890 --> 00:13:45,080
And so that's what you bid.

305
00:13:45,080 --> 00:13:46,290
OK.

306
00:13:46,290 --> 00:13:49,494
Now before I go on, let me
ask you one other question.

307
00:13:49,494 --> 00:13:51,660
Do you have a preference
for which color you'd pick?

308
00:13:54,480 --> 00:13:56,280
No preference?

309
00:13:56,280 --> 00:13:57,600
You sure?

310
00:13:57,600 --> 00:13:59,550
Why not?

311
00:13:59,550 --> 00:14:02,020
No preference?

312
00:14:02,020 --> 00:14:02,820
AUDIENCE: It's 50%.

313
00:14:02,820 --> 00:14:04,320
ANDREW LO: It's
50/50, that's right.

314
00:14:04,320 --> 00:14:06,090
So there's no
reason that you have

315
00:14:06,090 --> 00:14:07,710
a preference one or the other.

316
00:14:07,710 --> 00:14:09,480
However, I should
just let you know

317
00:14:09,480 --> 00:14:15,080
that in populations where
they do this kind of survey,

318
00:14:15,080 --> 00:14:17,820
there is a slight
preference for the color red

319
00:14:17,820 --> 00:14:19,479
in the population.

320
00:14:19,479 --> 00:14:20,520
It's kind of interesting.

321
00:14:20,520 --> 00:14:22,680
But, you know, all
right, fine, 50/50.

322
00:14:22,680 --> 00:14:23,970
You understand the game.

323
00:14:23,970 --> 00:14:26,040
There's no preference.

324
00:14:26,040 --> 00:14:29,220
OK, now, we're going to
play the exact same game

325
00:14:29,220 --> 00:14:35,230
except with urn B. Urn
B is identical to urn

326
00:14:35,230 --> 00:14:39,480
A in every respect except
that, in this case,

327
00:14:39,480 --> 00:14:42,050
I'm not going to tell
you the proportion.

328
00:14:42,050 --> 00:14:45,630
There could be
50/50 red and black,

329
00:14:45,630 --> 00:14:54,420
but it could be 40/60, or
70/30, or 85/15, or 100/0.

330
00:14:54,420 --> 00:14:59,040
I will promise you that
there are exactly 100 balls

331
00:14:59,040 --> 00:15:02,580
and there are, at most, two
colors, red and/or black,

332
00:15:02,580 --> 00:15:07,090
but beyond that, I will
not tell you anything.

333
00:15:07,090 --> 00:15:10,140
And now we're going to
play the exact same game.

334
00:15:10,140 --> 00:15:11,020
OK?

335
00:15:11,020 --> 00:15:14,170
So the question is,
first of all, what color

336
00:15:14,170 --> 00:15:15,840
would you prefer?

337
00:15:15,840 --> 00:15:18,360
Any of you have any
particular preference here?

338
00:15:18,360 --> 00:15:19,200
Why not?

339
00:15:19,200 --> 00:15:19,896
AUDIENCE: Red.

340
00:15:19,896 --> 00:15:20,520
ANDREW LO: Red.

341
00:15:20,520 --> 00:15:22,170
You'd prefer red?

342
00:15:22,170 --> 00:15:22,890
OK, yeah.

343
00:15:22,890 --> 00:15:23,910
AUDIENCE: Go for
black because you know

344
00:15:23,910 --> 00:15:25,196
that most people prefer red.

345
00:15:25,196 --> 00:15:27,570
ANDREW LO: Well, no, I didn't
say most people prefer red.

346
00:15:27,570 --> 00:15:29,005
I said there was a slight
preference for red,

347
00:15:29,005 --> 00:15:31,470
and, Ingrid, you seem to
fall into that category.

348
00:15:31,470 --> 00:15:33,050
And so you'd prefer
black instead.

349
00:15:33,050 --> 00:15:33,550
All right.

350
00:15:33,550 --> 00:15:36,081
Well, we're going to come
back to the color later on.

351
00:15:36,081 --> 00:15:37,830
But now let me ask
you, how much would you

352
00:15:37,830 --> 00:15:39,600
pay to play this game with me?

353
00:15:39,600 --> 00:15:41,110
Everything else is the same.

354
00:15:41,110 --> 00:15:45,120
$10,000 prize if you get
the color that I pick.

355
00:15:45,120 --> 00:15:45,927
All right?

356
00:15:45,927 --> 00:15:47,010
AUDIENCE: Do we trust you?

357
00:15:47,010 --> 00:15:48,306
ANDREW LO: What's that?

358
00:15:48,306 --> 00:15:49,620
AUDIENCE: Can we trust you?

359
00:15:49,620 --> 00:15:50,010
ANDREW LO: I don't know.

360
00:15:50,010 --> 00:15:50,722
Do you?

361
00:15:50,722 --> 00:15:51,650
AUDIENCE: [LAUGHTER]

362
00:15:51,650 --> 00:15:53,900
ANDREW LO: Well, first of
all, what is there to trust?

363
00:15:53,900 --> 00:15:55,399
AUDIENCE: Is there
really a hundred?

364
00:15:55,399 --> 00:15:57,480
ANDREW LO: I, yes,
that I will guarantee.

365
00:15:57,480 --> 00:16:00,750
So after the fact, that we will
verify that there is exactly

366
00:16:00,750 --> 00:16:03,510
100 balls and it
will be, at most,

367
00:16:03,510 --> 00:16:06,000
two colors, red and/or
black, and that's it.

368
00:16:06,000 --> 00:16:08,171
No other elements
will be introduced.

369
00:16:08,171 --> 00:16:08,670
OK?

370
00:16:08,670 --> 00:16:11,880
So that's perfectly verifiable,
and we will verify that

371
00:16:11,880 --> 00:16:13,440
after the fact,
that all prices are

372
00:16:13,440 --> 00:16:15,870
contingent upon
that verification.

373
00:16:15,870 --> 00:16:18,720
But that's the extent to
which you need to trust me.

374
00:16:18,720 --> 00:16:21,210
If you're asking
whether or not I'm

375
00:16:21,210 --> 00:16:24,210
the one that's picking the
distribution, then, yes, I am.

376
00:16:24,210 --> 00:16:25,810
I'm picking the distribution.

377
00:16:25,810 --> 00:16:27,420
But I don't
communicate with you,

378
00:16:27,420 --> 00:16:29,820
and you don't communicate
with me ahead of time,

379
00:16:29,820 --> 00:16:33,160
other than what I know
about you in this class

380
00:16:33,160 --> 00:16:35,370
and over the last semester.

381
00:16:35,370 --> 00:16:36,630
OK?

382
00:16:36,630 --> 00:16:38,610
So now, let's
figure out how much

383
00:16:38,610 --> 00:16:41,096
this will draw in
terms of the price.

384
00:16:41,096 --> 00:16:43,470
How many people will be willing
to play this game with me

385
00:16:43,470 --> 00:16:46,150
once for $1?

386
00:16:46,150 --> 00:16:48,580
$1,000?

387
00:16:48,580 --> 00:16:50,650
$2,000?

388
00:16:50,650 --> 00:16:52,040
$3,000?

389
00:16:52,040 --> 00:16:54,190
$4,000?

390
00:16:54,190 --> 00:16:56,890
$4,999?

391
00:16:56,890 --> 00:16:59,320
$5,010?

392
00:16:59,320 --> 00:17:00,460
OK.

393
00:17:00,460 --> 00:17:01,280
So it's about the--

394
00:17:01,280 --> 00:17:03,522
You'd pay more?

395
00:17:03,522 --> 00:17:05,730
You'd pay more to play this
game than the one before?

396
00:17:08,390 --> 00:17:11,099
I'm going to come back
to that in a minute.

397
00:17:11,099 --> 00:17:14,290
That's worth exploring.

398
00:17:14,290 --> 00:17:16,390
But before I get back
to Davide, Justin,

399
00:17:16,390 --> 00:17:19,500
you were willing to
pay the same for both.

400
00:17:19,500 --> 00:17:20,364
Why?

401
00:17:20,364 --> 00:17:22,497
AUDIENCE: Well,
because the proportion

402
00:17:22,497 --> 00:17:25,109
doesn't matter because
you pick your color,

403
00:17:25,109 --> 00:17:27,570
so it's kind of
out of your hands.

404
00:17:27,570 --> 00:17:29,850
You know it's either one
color or the other color.

405
00:17:29,850 --> 00:17:31,628
You know the
proportion's already

406
00:17:31,628 --> 00:17:34,216
been decided upon
so it should be

407
00:17:34,216 --> 00:17:35,850
the same as the previous game.

408
00:17:35,850 --> 00:17:38,710
ANDREW LO: OK, so you think
that the odds are the same

409
00:17:38,710 --> 00:17:41,050
in this case as before?

410
00:17:41,050 --> 00:17:43,832
Even though I'm the one
that picks the odds?

411
00:17:43,832 --> 00:17:46,664
AUDIENCE: When you
pick it beforehand

412
00:17:46,664 --> 00:17:48,674
and, I guess if it was random--

413
00:17:48,674 --> 00:17:49,840
ANDREW LO: Well, no, no, no.

414
00:17:49,840 --> 00:17:51,470
It's exactly what
I just told you.

415
00:17:51,470 --> 00:17:54,475
I get to pick the odds so,
unless you consider me random--

416
00:17:54,475 --> 00:17:56,600
AUDIENCE: Yeah, but we
still get to pick the ball--

417
00:17:56,600 --> 00:17:57,790
ANDREW LO: You get
to pick the ball.

418
00:17:57,790 --> 00:17:59,623
You have to pick the
color, I pick the ball.

419
00:17:59,623 --> 00:18:02,430
AUDIENCE: Yeah, I still
think it's about 50/50.

420
00:18:02,430 --> 00:18:03,300
ANDREW LO: 50/50.

421
00:18:03,300 --> 00:18:03,840
OK.

422
00:18:03,840 --> 00:18:04,060
Mike?

423
00:18:04,060 --> 00:18:06,710
AUDIENCE: Do you pick the odds
after we tell you our thoughts?

424
00:18:06,710 --> 00:18:09,971
ANDREW LO: No, it's clearly not.

425
00:18:09,971 --> 00:18:10,470
Yeah.

426
00:18:10,470 --> 00:18:11,260
No, no.

427
00:18:11,260 --> 00:18:12,580
That's a good point.

428
00:18:12,580 --> 00:18:14,840
And, no, we won't
cheat, in the sense

429
00:18:14,840 --> 00:18:17,930
that I will not pick the
distribution after you tell me

430
00:18:17,930 --> 00:18:20,220
what your color is.

431
00:18:20,220 --> 00:18:20,765
OK.

432
00:18:20,765 --> 00:18:21,590
Dannon?

433
00:18:21,590 --> 00:18:23,715
AUDIENCE: [INAUDIBLE] I
mean, you really don't know

434
00:18:23,715 --> 00:18:24,820
the probability, so--

435
00:18:24,820 --> 00:18:25,489
ANDREW LO: Yeah.

436
00:18:25,489 --> 00:18:27,530
AUDIENCE: You really need
to have one black ball?

437
00:18:27,530 --> 00:18:30,321
Since you're drawing,
you know, one more person

438
00:18:30,321 --> 00:18:31,446
who has to match the color.

439
00:18:31,446 --> 00:18:32,892
Then you have another
99 that are left.

440
00:18:32,892 --> 00:18:35,058
So your probability's are
really high at that point.

441
00:18:35,058 --> 00:18:37,595
Unless you're very unlucky,
with that very first ball

442
00:18:37,595 --> 00:18:39,340
as being black, then--

443
00:18:39,340 --> 00:18:41,647
So, but it could go either way.

444
00:18:41,647 --> 00:18:42,230
ANDREW LO: OK.

445
00:18:42,230 --> 00:18:44,320
So I'm-- So what's
the bottom line?

446
00:18:44,320 --> 00:18:47,337
Do you think it's the same
as before, or it's different?

447
00:18:47,337 --> 00:18:49,670
Would you bid the same amount
to play this game with me?

448
00:18:49,670 --> 00:18:50,620
AUDIENCE: I ended up
bidding the same amount.

449
00:18:50,620 --> 00:18:51,411
ANDREW LO: You did?

450
00:18:51,411 --> 00:18:51,940
OK.

451
00:18:51,940 --> 00:18:54,040
Anybody disagree with that?

452
00:18:54,040 --> 00:18:55,660
Anybody that would
not pay the same?

453
00:18:55,660 --> 00:18:57,430
A lot of you didn't
pay the same.

454
00:18:57,430 --> 00:18:58,554
Yeah, Jen?

455
00:18:58,554 --> 00:19:00,762
AUDIENCE: Well, I would just
calculate that the worst

456
00:19:00,762 --> 00:19:02,072
outcome here is worse.

457
00:19:02,072 --> 00:19:03,020
It's like $100.

458
00:19:03,020 --> 00:19:04,270
ANDREW LO: Yeah, that's right.

459
00:19:04,270 --> 00:19:05,396
AUDIENCE: I would pay $100.

460
00:19:05,396 --> 00:19:06,936
ANDREW LO: That is
the worst outcome.

461
00:19:06,936 --> 00:19:09,640
I mean, if you pick red, I could
have 100 black balls in there,

462
00:19:09,640 --> 00:19:12,138
and then you're going
to lose, no matter what.

463
00:19:12,138 --> 00:19:13,596
AUDIENCE: But that's
also the best.

464
00:19:13,596 --> 00:19:14,096
Yeah.

465
00:19:14,096 --> 00:19:14,900
Pick black.

466
00:19:14,900 --> 00:19:16,120
Because it's-- [INAUDIBLE]

467
00:19:16,120 --> 00:19:17,360
ANDREW LO: That's true, yes.

468
00:19:17,360 --> 00:19:18,901
If you look at it
the other way, it's

469
00:19:18,901 --> 00:19:22,170
the best case if you
did the opposite.

470
00:19:22,170 --> 00:19:22,780
Hmm.

471
00:19:22,780 --> 00:19:25,080
So what do we make of that?

472
00:19:25,080 --> 00:19:25,880
Yeah?

473
00:19:25,880 --> 00:19:27,270
AUDIENCE: I think
basically it's,

474
00:19:27,270 --> 00:19:29,290
the laws of standard
probability are the same.

475
00:19:29,290 --> 00:19:29,956
ANDREW LO: Yeah.

476
00:19:29,956 --> 00:19:31,955
AUDIENCE: Because, in
this case, the probability

477
00:19:31,955 --> 00:19:34,920
is you all being able
to pick the right ball.

478
00:19:34,920 --> 00:19:35,962
And that's 50/50.

479
00:19:35,962 --> 00:19:36,920
The right color, right?

480
00:19:36,920 --> 00:19:37,628
ANDREW LO: Is it?

481
00:19:37,628 --> 00:19:38,920
Is it 50/50 though?

482
00:19:38,920 --> 00:19:39,850
That's the question.

483
00:19:39,850 --> 00:19:40,360
Yeah.

484
00:19:40,360 --> 00:19:42,919
I'm picking the distribution,
you're picking the ball, right?

485
00:19:42,919 --> 00:19:43,460
Yeah, Davide?

486
00:19:43,460 --> 00:19:44,700
AUDIENCE: I'm not sure
I understand exactly.

487
00:19:44,700 --> 00:19:45,440
So--

488
00:19:45,440 --> 00:19:48,980
AUDIENCE: [LAUGHTER]

489
00:19:48,980 --> 00:19:50,720
ANDREW LO: That would
explain, certainly.

490
00:19:50,720 --> 00:19:52,271
Sure.

491
00:19:52,271 --> 00:19:52,770
Go ahead.

492
00:19:52,770 --> 00:19:54,220
AUDIENCE: Say the real
proportion is 50/50,

493
00:19:54,220 --> 00:19:55,390
I would pay you $5,000.

494
00:19:55,390 --> 00:19:56,105
ANDREW LO: Sure.

495
00:19:56,105 --> 00:19:58,490
AUDIENCE: If the real
proportion is zero,

496
00:19:58,490 --> 00:19:59,855
I wouldn't pay you any more.

497
00:19:59,855 --> 00:20:02,451
If the real proportion is one,
I wouldn't pay you any more.

498
00:20:02,451 --> 00:20:03,700
ANDREW LO: Yeah, that's right.

499
00:20:03,700 --> 00:20:04,200
Sure.

500
00:20:04,200 --> 00:20:05,860
If you knew that
this was-- if you

501
00:20:05,860 --> 00:20:08,890
had an insider's view of
what the proportion is,

502
00:20:08,890 --> 00:20:09,760
you might pay more.

503
00:20:09,760 --> 00:20:11,680
AUDIENCE: So, in other words,
I would pay more than $5,000.

504
00:20:11,680 --> 00:20:13,650
ANDREW LO: OK, but do you
think you know something then?

505
00:20:13,650 --> 00:20:15,502
AUDIENCE: No, but
I would say that I

506
00:20:15,502 --> 00:20:18,005
would pay more than $5,000.

507
00:20:18,005 --> 00:20:18,630
ANDREW LO: Why?

508
00:20:18,630 --> 00:20:19,129
Why?

509
00:20:19,129 --> 00:20:20,680
Only if you had
more information.

510
00:20:20,680 --> 00:20:21,405
But do you?

511
00:20:21,405 --> 00:20:23,030
Do you think you have
more information?

512
00:20:23,030 --> 00:20:24,870
AUDIENCE: Even if
it's less than 50/50,

513
00:20:24,870 --> 00:20:26,720
then I would pay
more than $5,000.

514
00:20:26,720 --> 00:20:29,721
ANDREW LO: But what if
it's more than 50/50?

515
00:20:29,721 --> 00:20:31,345
What if it's more
than, in other words,

516
00:20:31,345 --> 00:20:36,440
what if it's 60/40 against you,
as opposed to 40/60 with you?

517
00:20:36,440 --> 00:20:37,870
AUDIENCE: Because at 60/40--

518
00:20:37,870 --> 00:20:39,100
I mean, it's like--

519
00:20:39,100 --> 00:20:42,801
as if it were 100/0,
then I would be winning.

520
00:20:42,801 --> 00:20:43,300
Sure.

521
00:20:43,300 --> 00:20:44,883
ANDREW LO: That
depends on your color.

522
00:20:44,883 --> 00:20:46,180
It depends on your color.

523
00:20:46,180 --> 00:20:47,799
Yeah, David?

524
00:20:47,799 --> 00:20:50,400
AUDIENCE: If there's no
expectation that you're biased,

525
00:20:50,400 --> 00:20:52,984
you can't make that expectation
and the expectation of how

526
00:20:52,984 --> 00:20:54,960
much you could get is the same.

527
00:20:54,960 --> 00:20:55,735
ANDREW LO: OK.

528
00:20:55,735 --> 00:20:57,960
AUDIENCE: So if you're going
just on the expectation

529
00:20:57,960 --> 00:21:00,030
of what your bias is--

530
00:21:00,030 --> 00:21:00,910
ANDREW LO: Right.

531
00:21:00,910 --> 00:21:03,880
Well, so you guys have seen
me over the last 12 weeks.

532
00:21:03,880 --> 00:21:05,800
You know what my biases are.

533
00:21:05,800 --> 00:21:08,260
In this case, I've
already told you

534
00:21:08,260 --> 00:21:09,951
that I have some information.

535
00:21:09,951 --> 00:21:10,450
Right?

536
00:21:10,450 --> 00:21:14,200
That people, on average,
prefer red a little bit more

537
00:21:14,200 --> 00:21:15,220
than black.

538
00:21:15,220 --> 00:21:18,220
So actually, you have
that piece of information.

539
00:21:18,220 --> 00:21:19,510
What would you do with it?

540
00:21:19,510 --> 00:21:22,965
Does that tell you anything?

541
00:21:22,965 --> 00:21:24,340
AUDIENCE: It tells
us everything.

542
00:21:24,340 --> 00:21:26,480
ANDREW LO: Well, it
does tell you something.

543
00:21:26,480 --> 00:21:29,470
It tells you that I
know that people prefer

544
00:21:29,470 --> 00:21:32,020
red slightly more than black.

545
00:21:32,020 --> 00:21:38,290
But the problem with that
is that you know I know.

546
00:21:38,290 --> 00:21:40,780
And moreover, I know
you know I know,

547
00:21:40,780 --> 00:21:43,960
and you know I know
you know I know.

548
00:21:43,960 --> 00:21:47,710
When you take it back to
its infinite regression,

549
00:21:47,710 --> 00:21:49,710
you know what it comes up to?

550
00:21:49,710 --> 00:21:52,570
50/50.

551
00:21:52,570 --> 00:21:54,610
It turns out that,
mathematically,

552
00:21:54,610 --> 00:21:57,970
the odds for this
game are literally

553
00:21:57,970 --> 00:22:01,080
identical to the odds
of the previous game.

554
00:22:01,080 --> 00:22:05,460
But it's also the case that
when you take these two urns

555
00:22:05,460 --> 00:22:08,460
and you auction them
off to the public,

556
00:22:08,460 --> 00:22:11,640
there is a very, very
significant discount

557
00:22:11,640 --> 00:22:16,950
in what people will pay for urn
B. And we saw this in class.

558
00:22:16,950 --> 00:22:19,230
A lot fewer of you
raised your hands

559
00:22:19,230 --> 00:22:23,512
we were going to $3,000,
$4,000, and $4,999.

560
00:22:23,512 --> 00:22:24,220
And you know why?

561
00:22:24,220 --> 00:22:26,380
When you ask people,
why is it that you

562
00:22:26,380 --> 00:22:31,120
would pay less for this
urn than the previous one,

563
00:22:31,120 --> 00:22:35,400
they'll tell you very plainly
it's because, in this case,

564
00:22:35,400 --> 00:22:37,614
I don't know what the odds are.

565
00:22:37,614 --> 00:22:39,530
I know what the odds are
in the previous case,

566
00:22:39,530 --> 00:22:44,790
but I don't know what kind
of risks I'm dealing with.

567
00:22:44,790 --> 00:22:47,310
In most dictionaries
and thesauruses,

568
00:22:47,310 --> 00:22:50,560
risk and uncertainty
are considered synonyms.

569
00:22:50,560 --> 00:22:53,950
But in fact, we see here that,
from a decision-making point

570
00:22:53,950 --> 00:22:58,810
of view, when there is
uncertainty about your risk,

571
00:22:58,810 --> 00:23:02,170
you tend to shy away from
that because I don't know

572
00:23:02,170 --> 00:23:04,150
what I'm dealing with here.

573
00:23:04,150 --> 00:23:06,130
Now there is one
simple mechanism

574
00:23:06,130 --> 00:23:12,040
by which all of you in this
room could make it impossible

575
00:23:12,040 --> 00:23:17,020
for me to somehow
strategically pick the worst

576
00:23:17,020 --> 00:23:20,180
distribution against you.

577
00:23:20,180 --> 00:23:22,331
Anybody tell me what
that algorithm is?

578
00:23:22,331 --> 00:23:22,830
Yeah?

579
00:23:22,830 --> 00:23:24,330
AUDIENCE: If you
just flipped a coin

580
00:23:24,330 --> 00:23:25,910
to decide what color you pick?

581
00:23:25,910 --> 00:23:27,080
ANDREW LO: Exactly.

582
00:23:27,080 --> 00:23:31,610
If Brian flips his coin, not
my coin, if you take your coin

583
00:23:31,610 --> 00:23:35,730
and flip it and pick red or
black based upon the outcome,

584
00:23:35,730 --> 00:23:38,510
there's no way I could
possibly do anything

585
00:23:38,510 --> 00:23:40,520
to disadvantage you.

586
00:23:40,520 --> 00:23:43,040
Now, that's a real interesting
conclusion, isn't it?

587
00:23:43,040 --> 00:23:44,510
Because what that's
telling you is

588
00:23:44,510 --> 00:23:48,950
that by injecting
uncertainty into a system,

589
00:23:48,950 --> 00:23:53,030
you might actually make it
more preferable, more fair,

590
00:23:53,030 --> 00:23:55,537
by adding randomness in there.

591
00:23:55,537 --> 00:23:57,370
And yet when you tell
people that, they say,

592
00:23:57,370 --> 00:24:00,940
well, you know that sounds
great, I still don't want this.

593
00:24:00,940 --> 00:24:04,290
Because I don't want to
deal with the uncertainty.

594
00:24:04,290 --> 00:24:06,500
And that's hardwired.

595
00:24:06,500 --> 00:24:11,290
We as human beings shy
away from the unknown.

596
00:24:11,290 --> 00:24:14,740
OK, now, take this
example and apply it

597
00:24:14,740 --> 00:24:17,860
to a non-financial
corporation that

598
00:24:17,860 --> 00:24:21,420
has toxic assets as part
of its balance sheet

599
00:24:21,420 --> 00:24:24,160
where you have no idea what the
market value is, because there

600
00:24:24,160 --> 00:24:26,140
is no market, because
it's not trading.

601
00:24:26,140 --> 00:24:28,330
Because you don't know how
to evaluate the security

602
00:24:28,330 --> 00:24:31,030
because it contains too many
parts that you can't figure out

603
00:24:31,030 --> 00:24:34,150
what the legal,
never mind financial,

604
00:24:34,150 --> 00:24:37,189
ramifications are
for those assets.

605
00:24:37,189 --> 00:24:39,730
And you can understand now why
people say, oh, wait a minute,

606
00:24:39,730 --> 00:24:40,780
I don't want to play.

607
00:24:40,780 --> 00:24:43,330
I don't I don't care
what it's worth.

608
00:24:43,330 --> 00:24:44,410
I don't want to know.

609
00:24:44,410 --> 00:24:46,750
I don't want to deal
with it because I

610
00:24:46,750 --> 00:24:49,480
can't evaluate the risk.

611
00:24:49,480 --> 00:24:52,480
That's what's happening
in spades today.

612
00:24:52,480 --> 00:24:53,580
[? Shlomi? ?]

613
00:24:53,580 --> 00:24:54,960
AUDIENCE: In the
second game, you

614
00:24:54,960 --> 00:24:59,404
said that the people preferred
to pay less because they don't

615
00:24:59,404 --> 00:25:02,890
want to go with uncertainty.

616
00:25:02,890 --> 00:25:04,980
ANDREW LO: The uncertainty
about the risk.

617
00:25:04,980 --> 00:25:07,948
AUDIENCE: But uncertainty
about the risk is unimportant.

618
00:25:07,948 --> 00:25:08,947
ANDREW LO: That's right.

619
00:25:08,947 --> 00:25:10,655
AUDIENCE: They don't
understand the game.

620
00:25:10,655 --> 00:25:13,240
This is the reason why
they prefer to [INAUDIBLE]..

621
00:25:13,240 --> 00:25:16,030
ANDREW LO: Well, let
me put it this way.

622
00:25:16,030 --> 00:25:19,270
All of you here, I think,
are pretty sophisticated.

623
00:25:19,270 --> 00:25:22,090
All of you, by the way,
understand the mathematics

624
00:25:22,090 --> 00:25:23,320
of this game, I believe.

625
00:25:23,320 --> 00:25:23,860
Right?

626
00:25:23,860 --> 00:25:25,960
There's not that
many moving parts.

627
00:25:25,960 --> 00:25:29,500
If we were to do an
anonymous survey where

628
00:25:29,500 --> 00:25:31,540
I forced you to play with me--

629
00:25:31,540 --> 00:25:33,640
So, in other words,
I deducted $5,000

630
00:25:33,640 --> 00:25:35,679
from your payroll, so
you're playing with me,

631
00:25:35,679 --> 00:25:37,220
no matter whether
you like it or not.

632
00:25:37,220 --> 00:25:39,310
You just paid $5,000.

633
00:25:39,310 --> 00:25:45,630
But I did let you choose whether
you play with urn A or urn B.

634
00:25:45,630 --> 00:25:48,360
I bet you the result,
even in as sophisticated

635
00:25:48,360 --> 00:25:53,370
an audience as this, is that the
vast majority would pick urn A,

636
00:25:53,370 --> 00:25:55,650
not urn B. You don't
have to raise your hands.

637
00:25:55,650 --> 00:25:57,270
I won't put you on the spot.

638
00:25:57,270 --> 00:26:00,390
But my guess is that most
of you, if not all of you,

639
00:26:00,390 --> 00:26:03,630
would end up picking urn A
simply because you then can

640
00:26:03,630 --> 00:26:07,170
evaluate the uncertainty,
whereas with B, you can't.

641
00:26:07,170 --> 00:26:09,540
So you're right that,
mathematically, they're

642
00:26:09,540 --> 00:26:10,500
the same.

643
00:26:10,500 --> 00:26:13,500
But the fact is that we
shy away from anything

644
00:26:13,500 --> 00:26:16,110
we don't understand, and
so what that suggests

645
00:26:16,110 --> 00:26:19,264
is that there are
layers of risk.

646
00:26:19,264 --> 00:26:21,180
In fact, there was an
economist many years ago

647
00:26:21,180 --> 00:26:24,450
by the name of Frank
Knight who was trying

648
00:26:24,450 --> 00:26:28,710
to explain how it was that
certain people in the economy

649
00:26:28,710 --> 00:26:30,730
were so much richer than others.

650
00:26:30,730 --> 00:26:33,630
In fact, he was trying to
explain the great wealth

651
00:26:33,630 --> 00:26:35,880
of the big mega industrialists.

652
00:26:35,880 --> 00:26:39,300
These are the folks like
Pierpont Morgan who was worth,

653
00:26:39,300 --> 00:26:42,420
at that time, $100
million, which

654
00:26:42,420 --> 00:26:44,160
was an extraordinary
amount of wealth.

655
00:26:44,160 --> 00:26:48,030
The modern day version
today is Bill Gates.

656
00:26:48,030 --> 00:26:51,330
The question is,
should Bill Gates

657
00:26:51,330 --> 00:26:55,440
be worth $40 billion
or $50 billion?

658
00:26:55,440 --> 00:26:57,540
Does that make sense?

659
00:26:57,540 --> 00:27:00,480
I mean, is he really
that much more valuable

660
00:27:00,480 --> 00:27:02,910
in any way than any of us?

661
00:27:02,910 --> 00:27:06,050
Why should somebody like that
be worth so much more money?

662
00:27:06,050 --> 00:27:08,960
And Frank Knight
struggled with this

663
00:27:08,960 --> 00:27:10,910
because he didn't
understand how,

664
00:27:10,910 --> 00:27:14,340
from an economic perspective,
you could justify it.

665
00:27:14,340 --> 00:27:17,910
And he finally came up with
the following hypothesis

666
00:27:17,910 --> 00:27:22,050
or proposal for why
these entrepreneurs

667
00:27:22,050 --> 00:27:24,180
make so much money.

668
00:27:24,180 --> 00:27:29,070
It's because they are taking
on a kind of randomness that

669
00:27:29,070 --> 00:27:31,132
is not quantifiable.

670
00:27:31,132 --> 00:27:33,090
In other words, there's
two kinds of randomness

671
00:27:33,090 --> 00:27:34,620
in Frank Knight's perspective.

672
00:27:34,620 --> 00:27:36,580
One is risk.

673
00:27:36,580 --> 00:27:38,660
Risk is quantifiable.

674
00:27:38,660 --> 00:27:40,624
For example, mortality rates.

675
00:27:40,624 --> 00:27:43,040
I don't know how many of you
have looked at life insurance

676
00:27:43,040 --> 00:27:43,950
policies.

677
00:27:43,950 --> 00:27:47,492
But if you have, as I did when
I first had my first child,

678
00:27:47,492 --> 00:27:49,700
I decided to learn a little
bit about life insurance,

679
00:27:49,700 --> 00:27:53,330
and read mortality tables, and
looked at life expectancies,

680
00:27:53,330 --> 00:27:54,770
and so on.

681
00:27:54,770 --> 00:27:57,530
And I found it
enormously depressing

682
00:27:57,530 --> 00:28:02,390
because it turns
out that, if you

683
00:28:02,390 --> 00:28:10,880
tell me 15 facts about your
life, with those 15 facts

684
00:28:10,880 --> 00:28:13,220
I can actually
pinpoint your life

685
00:28:13,220 --> 00:28:17,420
expectancy to within
plus or minus 5 years,

686
00:28:17,420 --> 00:28:20,410
with an enormous
degree of accuracy.

687
00:28:20,410 --> 00:28:23,220
And, as far as I'm concerned,
that's much more information

688
00:28:23,220 --> 00:28:24,900
than I need.

689
00:28:24,900 --> 00:28:27,245
I don't need to know
when I'm going to die.

690
00:28:27,245 --> 00:28:27,870
Think about it.

691
00:28:27,870 --> 00:28:30,360
Actually, as economists
we always say,

692
00:28:30,360 --> 00:28:32,590
the more information
you have, the better.

693
00:28:32,590 --> 00:28:34,040
Do you really believe that?

694
00:28:34,040 --> 00:28:35,540
Do you want to know when
you're going to die?

695
00:28:35,540 --> 00:28:37,040
If I had that
information, would you

696
00:28:37,040 --> 00:28:39,320
want me to tell you when
you're going to die?

697
00:28:39,320 --> 00:28:41,230
The specific day and time?

698
00:28:41,230 --> 00:28:42,710
Would you like to know?

699
00:28:42,710 --> 00:28:44,030
Think about that.

700
00:28:44,030 --> 00:28:46,280
Well, it turns out with
these mortality tables,

701
00:28:46,280 --> 00:28:48,260
whether or not
you've ever smoked,

702
00:28:48,260 --> 00:28:51,170
do you drink, what kind of car
do you drive, have you ever

703
00:28:51,170 --> 00:28:52,130
had an accident?

704
00:28:52,130 --> 00:28:54,410
There is these 15
or 20 questions

705
00:28:54,410 --> 00:28:57,600
that, once you answer them,
you get put into a box.

706
00:28:57,600 --> 00:29:00,560
And that box has a life
expectancy that ends up

707
00:29:00,560 --> 00:29:03,060
being pretty darned accurate.

708
00:29:03,060 --> 00:29:06,980
Now, the point of this is that
the life insurance business

709
00:29:06,980 --> 00:29:09,410
is not currently
a growth business.

710
00:29:09,410 --> 00:29:11,360
I don't think
anybody here is dying

711
00:29:11,360 --> 00:29:15,140
to graduate from Sloan to
start a life insurance company.

712
00:29:15,140 --> 00:29:15,870
Why is that?

713
00:29:15,870 --> 00:29:19,310
It's because there's not
that much profit left,

714
00:29:19,310 --> 00:29:22,310
given that it's relatively
well understood how

715
00:29:22,310 --> 00:29:24,549
to forecast mortality rates.

716
00:29:24,549 --> 00:29:26,840
And once you forecast them,
you can price the insurance

717
00:29:26,840 --> 00:29:29,870
in such a way that you'll
be slightly profitable,

718
00:29:29,870 --> 00:29:32,900
and that's pretty much
all you're going to do.

719
00:29:32,900 --> 00:29:38,540
So the profit opportunities
for the kinds of randomness

720
00:29:38,540 --> 00:29:42,800
that you can define very
precisely, like mortality

721
00:29:42,800 --> 00:29:47,120
tables, or property
and casualty insurance,

722
00:29:47,120 --> 00:29:49,610
there's not a lot of
profit in that anymore.

723
00:29:49,610 --> 00:29:52,490
Maybe for the first person
who did it, yes, there was.

724
00:29:52,490 --> 00:29:54,440
But by that time,
at that time, there

725
00:29:54,440 --> 00:29:56,221
wasn't a lot of data on this.

726
00:29:56,221 --> 00:29:56,720
OK.

727
00:29:56,720 --> 00:29:58,340
So Knight said
that you don't get

728
00:29:58,340 --> 00:30:04,310
rewarded with untold sums of
wealth for taking on risk.

729
00:30:04,310 --> 00:30:09,080
What you get rewarded for
with untold sums of wealth

730
00:30:09,080 --> 00:30:12,050
is for taking on uncertainty.

731
00:30:12,050 --> 00:30:15,140
Uncertainty is what Frank
Knight called the risk

732
00:30:15,140 --> 00:30:17,720
that nobody can parameterize.

733
00:30:17,720 --> 00:30:19,354
So, nanotechnology.

734
00:30:19,354 --> 00:30:21,770
Who the heck knows whether or
not that's going to pay off?

735
00:30:21,770 --> 00:30:24,590
We don't even know what that
is, or whether it's really

736
00:30:24,590 --> 00:30:25,730
a business.

737
00:30:25,730 --> 00:30:29,570
That kind of uncertainty, if
you're willing to take that on,

738
00:30:29,570 --> 00:30:32,540
then there are those
who will make literally

739
00:30:32,540 --> 00:30:34,410
billions of dollars.

740
00:30:34,410 --> 00:30:37,220
And so the profits that
economists would argue

741
00:30:37,220 --> 00:30:39,200
have to be in
equilibrium, zero, right?

742
00:30:39,200 --> 00:30:41,120
Zero profit condition
is typically

743
00:30:41,120 --> 00:30:43,460
what an economic
equilibrium consists of.

744
00:30:43,460 --> 00:30:46,220
Zero profit only
applies to the risks

745
00:30:46,220 --> 00:30:49,960
that can be parameterized,
that can be well understood.

746
00:30:49,960 --> 00:30:54,640
And the point is that the
human tendency is to shy away

747
00:30:54,640 --> 00:30:58,090
from the uncertainties of life.

748
00:30:58,090 --> 00:31:00,970
We don't like things
we don't understand,

749
00:31:00,970 --> 00:31:03,050
and we will shrink from those.

750
00:31:03,050 --> 00:31:06,100
So if you are wired
in a different way

751
00:31:06,100 --> 00:31:08,830
to be willing to take
those kinds of risks,

752
00:31:08,830 --> 00:31:11,561
you're going to get paid
huge sums of money for that.

753
00:31:11,561 --> 00:31:12,060
Ike?

754
00:31:12,060 --> 00:31:15,614
AUDIENCE: In terms of the
zero sum math game, what about

755
00:31:15,614 --> 00:31:16,780
the silence of the cemetery?

756
00:31:16,780 --> 00:31:20,255
That Bill Gates is offset by
the thousands of entrepenuers

757
00:31:20,255 --> 00:31:22,330
who've failed and
are worth nothing?

758
00:31:22,330 --> 00:31:25,540
So, on average,
they're all mortal.

759
00:31:25,540 --> 00:31:28,300
ANDREW LO: Well, that may be
true that, on average, it all

760
00:31:28,300 --> 00:31:29,020
works out.

761
00:31:29,020 --> 00:31:32,560
But the point is that why do
you have such a huge disparity

762
00:31:32,560 --> 00:31:34,690
for certain kinds of
randomness whereas,

763
00:31:34,690 --> 00:31:37,250
for other kinds of randomness,
you have a nice bell curve?

764
00:31:37,250 --> 00:31:39,190
In other words, why
do you have this kind

765
00:31:39,190 --> 00:31:40,944
of tail risk going on?

766
00:31:40,944 --> 00:31:42,610
The tail risk that
you're talking about?

767
00:31:42,610 --> 00:31:43,900
The silence of the cemeteries.

768
00:31:43,900 --> 00:31:45,940
The fact that, for
every one Bill Gates,

769
00:31:45,940 --> 00:31:48,440
you've got several million
failed entrepreneurs,

770
00:31:48,440 --> 00:31:51,490
is because people are taking
on the kind of randomness

771
00:31:51,490 --> 00:31:53,530
that nobody knows
how to quantify.

772
00:31:53,530 --> 00:31:54,410
That's the argument.

773
00:31:54,410 --> 00:31:57,090
So it's consistent
with that perspective.

774
00:31:57,090 --> 00:31:57,880
OK?

775
00:31:57,880 --> 00:32:00,880
So that's why VCs
and entrepreneurs--

776
00:32:00,880 --> 00:32:03,070
those of you who want
to do your start-ups,

777
00:32:03,070 --> 00:32:04,420
I wish you great success.

778
00:32:04,420 --> 00:32:08,830
And in this classroom, I will
bet that at least one of you

779
00:32:08,830 --> 00:32:12,110
will be a billionaire
in 15 years.

780
00:32:12,110 --> 00:32:14,810
And I want to get to know you
very well between now and then.

781
00:32:17,360 --> 00:32:21,050
But the downside, which
all of you should know,

782
00:32:21,050 --> 00:32:24,380
the downside is that for the
rest of the entrepreneurs that

783
00:32:24,380 --> 00:32:28,760
don't make a billion
dollars, you'll make nothing.

784
00:32:28,760 --> 00:32:29,360
Zero.

785
00:32:29,360 --> 00:32:31,250
You'll lose everything.

786
00:32:31,250 --> 00:32:33,470
Now, that's a
reasonable trade-off

787
00:32:33,470 --> 00:32:38,480
if your particular
psychological makeup

788
00:32:38,480 --> 00:32:40,940
enjoys that kind of a thrill.

789
00:32:40,940 --> 00:32:42,680
But it's not for
everybody, and that's

790
00:32:42,680 --> 00:32:44,611
why people get paid
what they get paid.

791
00:32:44,611 --> 00:32:45,110
Right?

792
00:32:45,110 --> 00:32:49,130
So there is some logic
to this kind of puzzle.

793
00:32:49,130 --> 00:32:52,100
By the way, this
particular example

794
00:32:52,100 --> 00:32:54,080
was developed also
by a psychologist

795
00:32:54,080 --> 00:32:55,220
by the name of Ellsberg.

796
00:32:55,220 --> 00:32:57,770
And it's called the
Ellsberg paradox

797
00:32:57,770 --> 00:32:59,840
because this is a
situation where,

798
00:32:59,840 --> 00:33:02,090
despite the fact that
the odds are the same,

799
00:33:02,090 --> 00:33:04,370
people don't like
the uncertainties.

800
00:33:04,370 --> 00:33:08,240
And to me, and to a lot of
evolutionary biologists,

801
00:33:08,240 --> 00:33:11,030
and neurophysiologists,
this is not surprising.

802
00:33:11,030 --> 00:33:13,654
Because we are hardwired
to stay away from stuff

803
00:33:13,654 --> 00:33:15,320
we don't know because
it's the stuff you

804
00:33:15,320 --> 00:33:17,060
don't know that can kill you.

805
00:33:17,060 --> 00:33:18,890
And from an evolutionary
perspective,

806
00:33:18,890 --> 00:33:20,660
it makes a lot of sense.

807
00:33:20,660 --> 00:33:21,480
Question?

808
00:33:21,480 --> 00:33:24,175
AUDIENCE: I feel like a big part
of what got us in this whole

809
00:33:24,175 --> 00:33:27,498
mess is that [INAUDIBLE]
came unwired--

810
00:33:27,498 --> 00:33:29,860
People actually were chasing
stuff they didn't know

811
00:33:29,860 --> 00:33:31,342
and paying more for it.

812
00:33:31,342 --> 00:33:33,800
ANDREW LO: Well, I think it's
not so much they were chasing

813
00:33:33,800 --> 00:33:34,970
what they didn't
know but rather,

814
00:33:34,970 --> 00:33:37,344
they were chasing stuff that
they thought they knew that,

815
00:33:37,344 --> 00:33:38,371
in fact, they didn't.

816
00:33:38,371 --> 00:33:39,870
So we're going to
come back to that.

817
00:33:39,870 --> 00:33:42,770
Let me hold off on that
for just 15 minutes.

818
00:33:42,770 --> 00:33:45,410
I want to take that head
on when we actually look

819
00:33:45,410 --> 00:33:47,000
at the applications to finance.

820
00:33:47,000 --> 00:33:47,676
Zeke?

821
00:33:47,676 --> 00:33:49,759
AUDIENCE: I think there's
also like an [INAUDIBLE]

822
00:33:49,759 --> 00:33:51,007
version of this as well.

823
00:33:51,007 --> 00:33:52,090
It's because there's not--

824
00:33:52,090 --> 00:33:54,530
there's also not
doing your homework.

825
00:33:54,530 --> 00:33:56,695
Taking on some
uncertainty whereas you

826
00:33:56,695 --> 00:33:57,694
could be taking risk.

827
00:33:57,694 --> 00:33:58,360
ANDREW LO: Sure.

828
00:33:58,360 --> 00:33:59,360
Absolutely.

829
00:33:59,360 --> 00:34:04,220
It's possible that this kind
of risk is not systematic risk.

830
00:34:04,220 --> 00:34:05,960
And there's nothing
that says that you

831
00:34:05,960 --> 00:34:09,050
ought to get compensated for
that, except for the fact that,

832
00:34:09,050 --> 00:34:11,840
if you're talking
about very, very, very

833
00:34:11,840 --> 00:34:15,290
unusual kinds of
technologies, or businesses

834
00:34:15,290 --> 00:34:19,350
that nobody really understands
or is able to take on,

835
00:34:19,350 --> 00:34:21,020
that's where the
multibillion dollar

836
00:34:21,020 --> 00:34:23,000
paydays are going to come from.

837
00:34:23,000 --> 00:34:25,460
It's not going to come from
starting up a news stand,

838
00:34:25,460 --> 00:34:28,900
or setting up another
restaurant, or selling coffee.

839
00:34:28,900 --> 00:34:29,989
Well, coffee, maybe.

840
00:34:29,989 --> 00:34:33,590
Let me take that back because
we have some counter-examples.

841
00:34:33,590 --> 00:34:35,820
But the point is that,
if it's commoditized,

842
00:34:35,820 --> 00:34:37,699
if everybody understands
how to do it,

843
00:34:37,699 --> 00:34:41,290
then the profits ultimately are
going to get bid down to zero.

844
00:34:41,290 --> 00:34:41,840
OK.

845
00:34:41,840 --> 00:34:42,472
Yeah?

846
00:34:42,472 --> 00:34:44,664
AUDIENCE: It's also
brought to your attention,

847
00:34:44,664 --> 00:34:47,380
you know, Warren
Buffet's averages

848
00:34:47,380 --> 00:34:51,532
of the normal conventions
of, you know, [INAUDIBLE]..

849
00:34:51,532 --> 00:34:55,350
Then he says [INAUDIBLE]
standard of [INAUDIBLE]..

850
00:34:55,350 --> 00:34:55,850
Right?

851
00:34:55,850 --> 00:34:58,880
So that's, that's a huge, that's
a huge uncertainty factor.

852
00:34:58,880 --> 00:35:01,280
But yet he's still [INAUDIBLE].

853
00:35:01,280 --> 00:35:02,280
ANDREW LO: That's right.

854
00:35:02,280 --> 00:35:04,930
He may have made
money in other ways.

855
00:35:04,930 --> 00:35:08,380
Not necessarily by taking risks
that nobody else understood,

856
00:35:08,380 --> 00:35:10,510
but rather by being
early in taking

857
00:35:10,510 --> 00:35:13,330
advantage of opportunities
that other people didn't see.

858
00:35:13,330 --> 00:35:16,910
So there's more than one
way to make a lot of money.

859
00:35:16,910 --> 00:35:19,600
And we'll talk a bit about
that at the end, when

860
00:35:19,600 --> 00:35:23,440
I talk about how evolutionary
pressures ultimately allow

861
00:35:23,440 --> 00:35:25,360
businesses to fail and succeed.

862
00:35:25,360 --> 00:35:28,076
So I'll come back to that.

863
00:35:28,076 --> 00:35:29,950
Any other questions
about the example though?

864
00:35:29,950 --> 00:35:30,450
OK.

865
00:35:30,450 --> 00:35:33,880
So we've got two examples of
irrationalities that all of us

866
00:35:33,880 --> 00:35:35,500
seem to possess.

867
00:35:35,500 --> 00:35:37,540
And the third
example that I give

868
00:35:37,540 --> 00:35:39,080
is one that I think
you have seen,

869
00:35:39,080 --> 00:35:41,350
which is this video
clip of this gorilla.

870
00:35:41,350 --> 00:35:42,250
Right?

871
00:35:42,250 --> 00:35:45,950
This is a clear example
where we are all

872
00:35:45,950 --> 00:35:51,190
hardwired to see certain things
that we are looking to see,

873
00:35:51,190 --> 00:35:53,650
and when we're not
looking for other things,

874
00:35:53,650 --> 00:35:55,660
we don't see them at all.

875
00:35:55,660 --> 00:35:59,710
And this final example I give
because it provides, I think,

876
00:35:59,710 --> 00:36:05,560
incontrovertible evidence
that human cognitive abilities

877
00:36:05,560 --> 00:36:07,010
are limited.

878
00:36:07,010 --> 00:36:10,030
We are not infinitely
rational in the sense

879
00:36:10,030 --> 00:36:14,680
that we can compute,
observe, analyze, forecast,

880
00:36:14,680 --> 00:36:19,330
and remember, an infinite
amount of information.

881
00:36:19,330 --> 00:36:24,160
And so, after all of these
examples, this is why I said,

882
00:36:24,160 --> 00:36:26,860
you must think that the
human animal is the stupidest

883
00:36:26,860 --> 00:36:28,420
creature on earth.

884
00:36:28,420 --> 00:36:31,690
Because we suffer from all
of these behavioral biases.

885
00:36:31,690 --> 00:36:34,720
So the behaviorists
would conclude by saying,

886
00:36:34,720 --> 00:36:36,370
we rest our case.

887
00:36:36,370 --> 00:36:38,470
People are irrational,
therefore how

888
00:36:38,470 --> 00:36:42,130
can you rely on markets for
making any kind of decisions?

889
00:36:42,130 --> 00:36:47,410
Well, the official markets types
have a pretty powerful retort.

890
00:36:47,410 --> 00:36:49,630
And the powerful retort
comes in the form

891
00:36:49,630 --> 00:36:51,490
of what's known as the
Dutch Book theorem.

892
00:36:51,490 --> 00:36:53,440
This is an example
that I also give

893
00:36:53,440 --> 00:36:56,410
to try to illustrate
how it could be that,

894
00:36:56,410 --> 00:36:59,140
even though there are crazy
people some of the time,

895
00:36:59,140 --> 00:37:01,810
and certain people that
are crazy all of the time,

896
00:37:01,810 --> 00:37:04,840
that nevertheless the
smart money will win.

897
00:37:04,840 --> 00:37:06,400
So the example goes like this.

898
00:37:06,400 --> 00:37:09,880
You've got an event A that
we're going to bet on, OK?

899
00:37:09,880 --> 00:37:13,540
And the event is that the
S&P 500 falls by 5% or more

900
00:37:13,540 --> 00:37:15,257
next Monday.

901
00:37:15,257 --> 00:37:17,340
This used to be considered
a pretty extreme event.

902
00:37:17,340 --> 00:37:20,700
Now it's business
as usual, right?

903
00:37:20,700 --> 00:37:21,930
OK.

904
00:37:21,930 --> 00:37:25,770
So let's suppose that you, as
the irrational, behavioral,

905
00:37:25,770 --> 00:37:29,550
psychological creature that
behavioral economists would

906
00:37:29,550 --> 00:37:32,700
argue you are, you have the
following probability belief.

907
00:37:32,700 --> 00:37:36,270
The probability of A is 1/2
and the probability of NOT A

908
00:37:36,270 --> 00:37:38,890
is 3/4.

909
00:37:38,890 --> 00:37:42,000
Now some of you are chuckling
because you would never

910
00:37:42,000 --> 00:37:43,110
have such beliefs, right?

911
00:37:43,110 --> 00:37:45,690
Because the probabilities,
they don't add up to 1.

912
00:37:45,690 --> 00:37:48,180
Well, I guarantee each
and every one of you

913
00:37:48,180 --> 00:37:50,370
if I got you into
a darkly-lit room,

914
00:37:50,370 --> 00:37:53,100
and I asked you a series
of questions over time,

915
00:37:53,100 --> 00:37:56,670
you would ultimately end up
exhibiting biased probabilities

916
00:37:56,670 --> 00:37:59,430
just like this, because
the human cognitive

917
00:37:59,430 --> 00:38:03,360
faculties are not designed to
make correct, probabilistic

918
00:38:03,360 --> 00:38:03,990
inferences.

919
00:38:03,990 --> 00:38:06,210
This is not natural,
that we know

920
00:38:06,210 --> 00:38:09,360
how to make probabilistic,
just judgments.

921
00:38:09,360 --> 00:38:12,360
So let's suppose that
this is what you believe.

922
00:38:12,360 --> 00:38:14,460
Now, what's the problem
with this irrationality?

923
00:38:14,460 --> 00:38:16,140
Well, the problem
is that you would

924
00:38:16,140 --> 00:38:20,990
be willing to take both
of the following two bets.

925
00:38:20,990 --> 00:38:26,210
Bet B1 is a bet where if
A occurs, I pay you $1.

926
00:38:26,210 --> 00:38:28,880
If A doesn't occur,
you pay me $1.

927
00:38:28,880 --> 00:38:29,454
OK?

928
00:38:29,454 --> 00:38:31,370
And the reason you're
willing to take this bet

929
00:38:31,370 --> 00:38:36,740
is because the odds of
A occurring are 1/2.

930
00:38:36,740 --> 00:38:39,680
1/2 means 50/50, and so
you're willing to take

931
00:38:39,680 --> 00:38:43,750
those 50/50 odds with bet B1.

932
00:38:43,750 --> 00:38:45,970
However, because
you also believe

933
00:38:45,970 --> 00:38:49,280
that the probability
of not A is 3/4,

934
00:38:49,280 --> 00:38:51,250
which is one to
three odds, right?

935
00:38:51,250 --> 00:38:52,420
1/4 to 3/4.

936
00:38:52,420 --> 00:38:53,530
One to three odds.

937
00:38:53,530 --> 00:38:58,170
Because of that fact, you're
also willing to take bet B2.

938
00:38:58,170 --> 00:39:04,030
Bet B2 is a bet where I pay
you $1 if A doesn't occur,

939
00:39:04,030 --> 00:39:09,270
but you pay me $3
if A does occur.

940
00:39:09,270 --> 00:39:11,940
And again, the reason you
are willing to do that

941
00:39:11,940 --> 00:39:16,050
is because you believe that the
probability of A not occurring

942
00:39:16,050 --> 00:39:18,900
is 3/4.

943
00:39:18,900 --> 00:39:22,790
Now, if that's the case, you're
willing to take bet B1 and B2,

944
00:39:22,790 --> 00:39:24,770
here's what we're going to do.

945
00:39:24,770 --> 00:39:31,900
I'm going to place $50
on B1 and $25 on B2.

946
00:39:31,900 --> 00:39:36,450
That's the wager that I
want to engage in with you.

947
00:39:36,450 --> 00:39:39,660
If we do that, then
let's see what happens.

948
00:39:39,660 --> 00:39:46,140
If A occurs, I lose my $50 on
B1 but I win three to one odds

949
00:39:46,140 --> 00:39:49,890
on my $25 on B2,
so I win $75 on B2,

950
00:39:49,890 --> 00:39:55,380
so I make $25
profit if A occurs.

951
00:39:55,380 --> 00:39:58,400
If A doesn't occur,
what happens?

952
00:39:58,400 --> 00:40:02,820
Well, if A doesn't occur,
I lose my $25 on B2,

953
00:40:02,820 --> 00:40:06,100
but I make $50 on B1.

954
00:40:06,100 --> 00:40:10,560
I make $25 again.

955
00:40:10,560 --> 00:40:13,410
Heads I win, tails you lose.

956
00:40:13,410 --> 00:40:17,580
No matter what
happens, I make $25,

957
00:40:17,580 --> 00:40:20,460
so I'm going to like this a lot.

958
00:40:20,460 --> 00:40:24,390
And I'm going to keep doing
it, and doing it, and doing it,

959
00:40:24,390 --> 00:40:26,620
until one of two things happens.

960
00:40:26,620 --> 00:40:32,520
Either you run out of money,
or you change your probability

961
00:40:32,520 --> 00:40:34,930
beliefs.

962
00:40:34,930 --> 00:40:39,220
And it turns out that there
is only one set of probability

963
00:40:39,220 --> 00:40:44,680
beliefs for which I cannot
construct this free lunch.

964
00:40:44,680 --> 00:40:46,460
You know that is?

965
00:40:46,460 --> 00:40:48,830
It's when the
probabilities sum to one.

966
00:40:48,830 --> 00:40:52,970
So the point is that there
is a limit to how irrational

967
00:40:52,970 --> 00:40:54,620
you can get.

968
00:40:54,620 --> 00:40:59,930
Because as long as I'm there
to take advantage of you,

969
00:40:59,930 --> 00:41:03,120
I'm simply going to pump money
out of you until you cry,

970
00:41:03,120 --> 00:41:03,620
uncle.

971
00:41:03,620 --> 00:41:06,370
Until you change your beliefs.

972
00:41:06,370 --> 00:41:06,870
Yeah?

973
00:41:06,870 --> 00:41:08,734
AUDIENCE: Are there any examples
of where this has actually

974
00:41:08,734 --> 00:41:09,200
been exploited?

975
00:41:09,200 --> 00:41:10,116
ANDREW LO: Absolutely.

976
00:41:10,116 --> 00:41:11,490
There are tons of examples.

977
00:41:11,490 --> 00:41:13,530
There are tons of
examples of mispricing.

978
00:41:13,530 --> 00:41:15,170
In fact, the example
that I gave you

979
00:41:15,170 --> 00:41:16,920
when we were doing
fixed income arbitrage,

980
00:41:16,920 --> 00:41:20,100
when we had these different
treasury securities

981
00:41:20,100 --> 00:41:24,300
that had different yields for
the very, very same maturities,

982
00:41:24,300 --> 00:41:27,450
and you could solve using
simultaneous linear equations

983
00:41:27,450 --> 00:41:28,740
for the arbitrage.

984
00:41:28,740 --> 00:41:32,150
That is exactly this example.

985
00:41:32,150 --> 00:41:36,490
So there are many, many
examples where this occurs.

986
00:41:36,490 --> 00:41:40,630
And it doesn't last,
typically, because people

987
00:41:40,630 --> 00:41:43,900
get tired of losing money,
and ultimately either they

988
00:41:43,900 --> 00:41:48,010
run out of that money,
or they get wise.

989
00:41:48,010 --> 00:41:51,190
So the idea behind this
argument that efficient markets

990
00:41:51,190 --> 00:41:55,150
folks make is that, sure, people
may be irrational from time

991
00:41:55,150 --> 00:41:56,950
to time, and there may
be some people that

992
00:41:56,950 --> 00:41:59,410
are always irrational,
but as long

993
00:41:59,410 --> 00:42:02,750
as there are smart people
out there, as long as there's

994
00:42:02,750 --> 00:42:05,920
smart money out there,
then it doesn't matter.

995
00:42:05,920 --> 00:42:10,000
Because the probabilities
will sum to one through market

996
00:42:10,000 --> 00:42:11,840
forces.

997
00:42:11,840 --> 00:42:16,520
And so then the question is, how
strong are those market forces?

998
00:42:16,520 --> 00:42:18,740
And I would point to
the current crisis

999
00:42:18,740 --> 00:42:22,100
to tell you that
those market forces,

1000
00:42:22,100 --> 00:42:26,690
while if they sound good on
paper and, at a place like MIT,

1001
00:42:26,690 --> 00:42:31,010
I promise you that what John
Maynard Keynes said decades ago

1002
00:42:31,010 --> 00:42:32,270
is still true.

1003
00:42:32,270 --> 00:42:36,380
Which is, the market can stay
irrational much, much longer

1004
00:42:36,380 --> 00:42:39,200
than you can stay solvent.

1005
00:42:39,200 --> 00:42:41,930
In other words, if you
are the only sane person

1006
00:42:41,930 --> 00:42:45,260
in an irrational world, you
are the one with the problem.

1007
00:42:45,260 --> 00:42:48,680
Because you can't take
advantage of these opportunities

1008
00:42:48,680 --> 00:42:55,510
before ultimately the forces
of irrationality overwhelm you.

1009
00:42:55,510 --> 00:42:58,470
At least that's what John
Maynard Keynes thought.

1010
00:42:58,470 --> 00:43:01,090
Now, it seems to
suggest then that it's

1011
00:43:01,090 --> 00:43:03,310
a question about magnitudes.

1012
00:43:03,310 --> 00:43:06,510
In other words, you've
got smart people here,

1013
00:43:06,510 --> 00:43:08,680
you've got irrational
people here--

1014
00:43:08,680 --> 00:43:15,020
They're going to interact and
may the bigger pocketbook win.

1015
00:43:15,020 --> 00:43:17,990
Question is, who's got
the bigger pocketbook?

1016
00:43:17,990 --> 00:43:21,350
Who's got the stronger
set of forces?

1017
00:43:21,350 --> 00:43:26,140
The irrational guys or the
efficient markets guys?

1018
00:43:26,140 --> 00:43:28,150
And before I answer
that question,

1019
00:43:28,150 --> 00:43:31,330
or at least propose an
answer to that question,

1020
00:43:31,330 --> 00:43:33,850
I need to now beg
your indulgence

1021
00:43:33,850 --> 00:43:36,610
for the next 15 or 20
minutes because we're

1022
00:43:36,610 --> 00:43:38,560
going to take a little
bit of a detour.

1023
00:43:38,560 --> 00:43:40,720
I want to bring in
some evidence from

1024
00:43:40,720 --> 00:43:44,230
the cognitive neurosciences
that will speak directly

1025
00:43:44,230 --> 00:43:49,030
to this issue of what we
even mean by rationality.

1026
00:43:49,030 --> 00:43:53,710
It turns out that what
economists mean by rational

1027
00:43:53,710 --> 00:43:59,590
is not what you and I take
to be rational from a daily,

1028
00:43:59,590 --> 00:44:03,060
kind of, activities
type of perspective.

1029
00:44:03,060 --> 00:44:05,510
So I want to tell you
about some research

1030
00:44:05,510 --> 00:44:08,980
in the cognitive
neurosciences that really

1031
00:44:08,980 --> 00:44:12,490
changes the way we
think about rationality

1032
00:44:12,490 --> 00:44:14,050
and decision-making.

1033
00:44:14,050 --> 00:44:18,070
To do that, I'm going to tell
you about a particular case

1034
00:44:18,070 --> 00:44:20,680
study that was described
in a book written

1035
00:44:20,680 --> 00:44:24,070
by a neuroscientist by the
name of Antonio Damasio.

1036
00:44:24,070 --> 00:44:27,520
Damasio is a
clinical neurologist,

1037
00:44:27,520 --> 00:44:30,880
meaning that he sees patients
as well as does research.

1038
00:44:30,880 --> 00:44:35,784
And about 15 years ago, he wrote
a book titled Descartes' Error,

1039
00:44:35,784 --> 00:44:37,450
and in a few minutes
I'll explain to you

1040
00:44:37,450 --> 00:44:39,100
where that title comes from.

1041
00:44:39,100 --> 00:44:42,160
But in this book, he
describes his experience

1042
00:44:42,160 --> 00:44:43,770
in dealing with one
of his patients,

1043
00:44:43,770 --> 00:44:46,900
codenamed "Elliot," who
developed a brain tumor that

1044
00:44:46,900 --> 00:44:48,250
had to be surgically removed.

1045
00:44:48,250 --> 00:44:51,430
And, as you may know, with
these kinds of surgeries,

1046
00:44:51,430 --> 00:44:54,220
you've got to remove a
healthy portion of the brain

1047
00:44:54,220 --> 00:44:55,690
as well as the
tumor to make sure

1048
00:44:55,690 --> 00:44:57,065
that you've got
all of the tumor,

1049
00:44:57,065 --> 00:45:00,460
and you don't leave
anything remaining.

1050
00:45:00,460 --> 00:45:03,080
So after they took
out this tumor,

1051
00:45:03,080 --> 00:45:05,830
and after Eliot recovered
from the surgery,

1052
00:45:05,830 --> 00:45:09,160
they gave him a battery of
tests to see whether or not

1053
00:45:09,160 --> 00:45:11,860
any of his cognitive
abilities were impaired

1054
00:45:11,860 --> 00:45:14,680
because of this surgery.

1055
00:45:14,680 --> 00:45:18,610
They gave him a perception
test, memory test, IQ test,

1056
00:45:18,610 --> 00:45:21,400
learning ability, language
ability, arithmetic.

1057
00:45:21,400 --> 00:45:23,980
With every single
test, Eliot passed

1058
00:45:23,980 --> 00:45:28,210
with flying colors, either
average or well above-average.

1059
00:45:28,210 --> 00:45:34,240
And yet, within weeks after
the recovery of his surgery,

1060
00:45:34,240 --> 00:45:37,870
he was fired from his
job, his wife left him,

1061
00:45:37,870 --> 00:45:39,620
and, for all intents
and purposes,

1062
00:45:39,620 --> 00:45:41,190
he had to be institutionalized.

1063
00:45:41,190 --> 00:45:44,350
He couldn't live on his own
because his behavior became

1064
00:45:44,350 --> 00:45:46,340
so irrational.

1065
00:45:46,340 --> 00:45:50,200
Now let me describe to you
what kind of irrationality

1066
00:45:50,200 --> 00:45:51,760
we're talking about.

1067
00:45:51,760 --> 00:45:58,150
Damasio describes what
Elliot did when he was

1068
00:45:58,150 --> 00:46:00,590
engaged in a task at his job.

1069
00:46:00,590 --> 00:46:02,500
He was an accountant
at the time.

1070
00:46:02,500 --> 00:46:05,290
And so Damasio writes,
"When the job called

1071
00:46:05,290 --> 00:46:08,200
for interrupting an activity
and turning to another,

1072
00:46:08,200 --> 00:46:10,010
he might persist
nonetheless, seemingly

1073
00:46:10,010 --> 00:46:11,260
losing sight of his main goal.

1074
00:46:11,260 --> 00:46:13,030
Or he might interrupt
the activity

1075
00:46:13,030 --> 00:46:14,530
he had engaged, to
turn to something

1076
00:46:14,530 --> 00:46:16,930
he found more captivating
at that particular moment.

1077
00:46:16,930 --> 00:46:19,870
The flow of work was stopped.

1078
00:46:19,870 --> 00:46:22,234
One might say that
the particular step

1079
00:46:22,234 --> 00:46:23,650
of the task at
which Elliot balked

1080
00:46:23,650 --> 00:46:25,790
was actually being
carried out too well,

1081
00:46:25,790 --> 00:46:28,690
and at the expense of
the overall purpose.

1082
00:46:28,690 --> 00:46:32,260
One might say that Eliot had
become irrational concerning

1083
00:46:32,260 --> 00:46:34,800
the larger frame of behavior."

1084
00:46:34,800 --> 00:46:41,200
Now, what Damasio was telling
you is a very simple example

1085
00:46:41,200 --> 00:46:43,210
that he gave in the
book, where Eliot

1086
00:46:43,210 --> 00:46:46,320
was asked to write a letter
to one of his clients.

1087
00:46:46,320 --> 00:46:49,680
He'd fire up the word
processor and, before he

1088
00:46:49,680 --> 00:46:51,990
started typing
that letter, he had

1089
00:46:51,990 --> 00:46:54,390
to pick just the right font.

1090
00:46:54,390 --> 00:46:57,550
And it took him three
hours to pick that font.

1091
00:46:57,550 --> 00:46:59,980
Three hours, literally.

1092
00:46:59,980 --> 00:47:02,740
Now, those of you who
use Microsoft Word,

1093
00:47:02,740 --> 00:47:04,172
I know that when
you first got it,

1094
00:47:04,172 --> 00:47:05,380
you did the same thing I did.

1095
00:47:05,380 --> 00:47:07,480
You pulled down that
font menu, and you

1096
00:47:07,480 --> 00:47:11,770
went through Arial, Sans
Serif, Times New Roman.

1097
00:47:11,770 --> 00:47:14,950
You checked out all those
fonts, but it probably

1098
00:47:14,950 --> 00:47:17,360
didn't take you three hours.

1099
00:47:17,360 --> 00:47:21,770
Moreover, it didn't take you
three hours every time you

1100
00:47:21,770 --> 00:47:23,790
had to write a letter.

1101
00:47:23,790 --> 00:47:26,250
Well, it did for Eliot
after the surgery.

1102
00:47:26,250 --> 00:47:31,020
Every letter he wrote took
three hours at the start

1103
00:47:31,020 --> 00:47:35,570
because he had to pick
just the right font.

1104
00:47:35,570 --> 00:47:42,050
And it didn't dawn on Damasio
until he gave Elliot one test

1105
00:47:42,050 --> 00:47:46,740
that he finally failed
as to what was going on.

1106
00:47:46,740 --> 00:47:48,580
And let me tell you
what the test was.

1107
00:47:48,580 --> 00:47:51,520
Psychologists
often use something

1108
00:47:51,520 --> 00:47:57,310
called eye-blink response rate
to measure emotional reaction.

1109
00:47:57,310 --> 00:47:59,860
It turns out that for
normal human beings, when

1110
00:47:59,860 --> 00:48:02,440
you have a strong
emotional response,

1111
00:48:02,440 --> 00:48:04,850
you blink your eyes a little
bit more quickly than usual.

1112
00:48:04,850 --> 00:48:05,350
All right?

1113
00:48:05,350 --> 00:48:08,020
Those of you who've ever
seen the Betty Boop cartoons

1114
00:48:08,020 --> 00:48:10,300
where they have a
female character trying

1115
00:48:10,300 --> 00:48:13,270
to flirt with somebody else,
that the way they'll indicate

1116
00:48:13,270 --> 00:48:16,750
flirtation is blinking their
eyes a little bit more quickly

1117
00:48:16,750 --> 00:48:18,130
or batting her eyelashes.

1118
00:48:18,130 --> 00:48:18,820
Right?

1119
00:48:18,820 --> 00:48:21,020
Those of you who are
serious poker players,

1120
00:48:21,020 --> 00:48:23,290
you'll know that
there's a poker tell,

1121
00:48:23,290 --> 00:48:26,380
meaning a physical
characteristic

1122
00:48:26,380 --> 00:48:29,752
that will signal a strong,
emotional reaction on the part

1123
00:48:29,752 --> 00:48:30,460
of your opponent.

1124
00:48:30,460 --> 00:48:32,590
And that is, when they
look at their hand,

1125
00:48:32,590 --> 00:48:34,769
they blink their eyes a
little bit more rapidly.

1126
00:48:34,769 --> 00:48:37,060
Either it's a really strong
hand or a really weak hand.

1127
00:48:37,060 --> 00:48:39,070
That's, by the way, why
many poker players--

1128
00:48:39,070 --> 00:48:40,900
they wear sunglasses
when they play.

1129
00:48:40,900 --> 00:48:45,890
Not to look cool but to
hide their reactions.

1130
00:48:45,890 --> 00:48:47,420
What they do is
they take a subject

1131
00:48:47,420 --> 00:48:49,211
and put them in front
of a computer screen.

1132
00:48:49,211 --> 00:48:50,660
Attach electrodes
to their eyelids

1133
00:48:50,660 --> 00:48:52,880
to measure how quickly
they're blinking their eyes,

1134
00:48:52,880 --> 00:48:54,650
and they show the
subject strong,

1135
00:48:54,650 --> 00:48:56,400
emotionally-charged images.

1136
00:48:56,400 --> 00:48:59,240
Things like
pornography or scenes

1137
00:48:59,240 --> 00:49:01,910
from Nazi concentration camps.

1138
00:49:01,910 --> 00:49:03,890
And when we have strong
emotional reactions,

1139
00:49:03,890 --> 00:49:07,730
you can measure the
increase in the rate

1140
00:49:07,730 --> 00:49:10,440
at which we blink our eyes.

1141
00:49:10,440 --> 00:49:15,530
It turns out that Elliot had
no change in his eye-blink

1142
00:49:15,530 --> 00:49:16,400
response rates.

1143
00:49:16,400 --> 00:49:20,310
None whatsoever.

1144
00:49:20,310 --> 00:49:22,200
And so Damasio
finally asked Elliot,

1145
00:49:22,200 --> 00:49:26,250
"Elliot, how do you feel?"

1146
00:49:26,250 --> 00:49:28,270
And Elliot said, "You
know, it's funny.

1147
00:49:28,270 --> 00:49:32,870
After the surgery, I
observed that the things

1148
00:49:32,870 --> 00:49:37,600
that I used to enjoy before
the surgery, I don't anymore.

1149
00:49:37,600 --> 00:49:40,810
For example, I used to like
steak dinners, red wine,

1150
00:49:40,810 --> 00:49:42,190
and Mozart.

1151
00:49:42,190 --> 00:49:45,930
And I've experienced all
of those things afterwards,

1152
00:49:45,930 --> 00:49:47,760
and I don't feel anything.

1153
00:49:47,760 --> 00:49:51,590
I know I should be
enjoying them, but I don't.

1154
00:49:51,590 --> 00:49:52,920
I don't feel anything.

1155
00:49:52,920 --> 00:49:54,290
I don't dislike it.

1156
00:49:54,290 --> 00:49:58,605
I just don't really enjoy it
as much as I know I should."

1157
00:49:58,605 --> 00:49:59,980
Now, I don't know
how many of you

1158
00:49:59,980 --> 00:50:01,810
have had that kind of reaction.

1159
00:50:01,810 --> 00:50:03,550
I have fairly recently.

1160
00:50:03,550 --> 00:50:06,670
My eight-year-old son and
I were at the supermarket

1161
00:50:06,670 --> 00:50:09,730
the other day, and I was
going by the candy aisle,

1162
00:50:09,730 --> 00:50:12,830
and I noticed that there were
Cracker Jacks on the shelf.

1163
00:50:12,830 --> 00:50:14,830
I don't know how many of
you know what they are?

1164
00:50:14,830 --> 00:50:18,970
Peanut-coated popcorn
candy and a prize, right?

1165
00:50:18,970 --> 00:50:21,490
Well, I loved Cracker Jacks
when I was eight years old.

1166
00:50:21,490 --> 00:50:24,370
I remember that very distinctly,
just really enjoying it.

1167
00:50:24,370 --> 00:50:27,250
So I decided I was going to
get some for my eight-year-old.

1168
00:50:27,250 --> 00:50:30,730
So I got some for him,
and I tried some as well.

1169
00:50:30,730 --> 00:50:34,390
And I remember thinking,
gee, I know I should really

1170
00:50:34,390 --> 00:50:36,430
be enjoying this, but I'm not.

1171
00:50:36,430 --> 00:50:39,310
It's not-- Maybe my tastes
have changed, I don't know,

1172
00:50:39,310 --> 00:50:43,240
but it's not nearly as wonderful
as I remembered it to be.

1173
00:50:43,240 --> 00:50:48,370
Imagine having that
feeling about everything.

1174
00:50:48,370 --> 00:50:53,770
All your fears, anxieties,
your great desires,

1175
00:50:53,770 --> 00:50:57,111
all of your emotions, gone.

1176
00:50:57,111 --> 00:50:58,110
You don't feel anything.

1177
00:50:58,110 --> 00:51:00,980
You sort of feel numb inside.

1178
00:51:00,980 --> 00:51:04,370
It turns out that
Damasio conjectured

1179
00:51:04,370 --> 00:51:08,120
that what we think of
as rational behavior

1180
00:51:08,120 --> 00:51:11,570
actually requires emotion.

1181
00:51:11,570 --> 00:51:15,260
In other words, you have
to be able to feel in order

1182
00:51:15,260 --> 00:51:17,870
to act rationally.

1183
00:51:17,870 --> 00:51:23,420
Now, that's really very
different from the typical kind

1184
00:51:23,420 --> 00:51:26,780
of impression where rationality
is at one end of the spectrum

1185
00:51:26,780 --> 00:51:29,414
and emotion, fear, and
greed are at the other end

1186
00:51:29,414 --> 00:51:30,080
of the spectrum.

1187
00:51:30,080 --> 00:51:32,820
This is Descartes'
perspective on the world.

1188
00:51:32,820 --> 00:51:34,820
This is why the book is
titled Descartes' Error.

1189
00:51:34,820 --> 00:51:37,790
He's arguing that
Descartes got it all wrong.

1190
00:51:37,790 --> 00:51:42,110
That, in fact, emotion is the
opposite side of the same coin.

1191
00:51:42,110 --> 00:51:46,660
If you can't feel, then
you can't be rational.

1192
00:51:46,660 --> 00:51:51,832
So this calls into question
what we think of as rationality.

1193
00:51:51,832 --> 00:51:54,040
And I'm going to beg your
indulgence for another five

1194
00:51:54,040 --> 00:51:58,120
minutes as I tell you more about
the neurophysiological basis

1195
00:51:58,120 --> 00:51:59,530
for this observation.

1196
00:51:59,530 --> 00:52:02,410
It turns out that the part of
the brain that they took out

1197
00:52:02,410 --> 00:52:06,010
from Elliot with his
tumor is the part that

1198
00:52:06,010 --> 00:52:09,620
was responsible for
emotional reaction.

1199
00:52:09,620 --> 00:52:12,910
And there have been
a number of theories

1200
00:52:12,910 --> 00:52:15,800
that have been developed now
about how the brain works.

1201
00:52:15,800 --> 00:52:19,420
So I want to describe probably
the most popular theory,

1202
00:52:19,420 --> 00:52:21,329
at least as of
about 10 years ago,

1203
00:52:21,329 --> 00:52:22,870
and you'll get a
better understanding

1204
00:52:22,870 --> 00:52:23,703
for what's going on.

1205
00:52:23,703 --> 00:52:27,430
And this is the part that I
think will change your lives.

1206
00:52:27,430 --> 00:52:31,580
So it turns out that the
brain, which is pictured here

1207
00:52:31,580 --> 00:52:36,370
in cross-section, is not
nearly the homogeneous organ

1208
00:52:36,370 --> 00:52:39,754
that we often take it to be.

1209
00:52:39,754 --> 00:52:41,420
It turns out that the
brain is comprised

1210
00:52:41,420 --> 00:52:42,740
of different components.

1211
00:52:42,740 --> 00:52:45,950
And a neuroscientist by
the name of Paul MacLean

1212
00:52:45,950 --> 00:52:48,200
about 10 years ago
came up with what's

1213
00:52:48,200 --> 00:52:51,260
now called the triune
model of the brain, which

1214
00:52:51,260 --> 00:52:55,980
decomposes the brain into
three separate components.

1215
00:52:55,980 --> 00:52:59,220
The first component is located
at the bottom of the brain.

1216
00:52:59,220 --> 00:53:01,550
It's also known
as the brain stem.

1217
00:53:01,550 --> 00:53:03,080
It's the part of
the brain that sits

1218
00:53:03,080 --> 00:53:05,670
on top of your
spinal cord, and all

1219
00:53:05,670 --> 00:53:08,840
of the nerves that go up
through your spinal cord

1220
00:53:08,840 --> 00:53:11,940
attaches into that brain stem.

1221
00:53:11,940 --> 00:53:15,030
And neuroscientists have
determined that the brain stem

1222
00:53:15,030 --> 00:53:18,870
is responsible for all of the,
sort of, bodily functions that

1223
00:53:18,870 --> 00:53:21,580
you take for granted, and
typically you cannot control,

1224
00:53:21,580 --> 00:53:23,670
unless you you're
a Zen Sufi master.

1225
00:53:23,670 --> 00:53:26,820
Things like heart
rate, breathing,

1226
00:53:26,820 --> 00:53:28,620
body temperature
regulation, things that we

1227
00:53:28,620 --> 00:53:30,600
need for basic survival.

1228
00:53:30,600 --> 00:53:33,510
And MacLean called the brain
stem the reptilian brain

1229
00:53:33,510 --> 00:53:35,250
because this is the
part of the brain

1230
00:53:35,250 --> 00:53:38,550
that we have in common with
virtually all vertebrates.

1231
00:53:38,550 --> 00:53:41,280
Reptiles, mammals,
anything with a spinal cord

1232
00:53:41,280 --> 00:53:44,580
will have a brain stem
sitting on top of that.

1233
00:53:44,580 --> 00:53:46,410
And from an evolutionary
perspective,

1234
00:53:46,410 --> 00:53:49,260
it is the oldest
part of the brain,

1235
00:53:49,260 --> 00:53:52,290
and the part that really
controls the very most

1236
00:53:52,290 --> 00:53:55,990
primitive functions of life.

1237
00:53:55,990 --> 00:53:58,290
The second part of the brain
that MacLean identified is

1238
00:53:58,290 --> 00:54:02,040
the middle part, known
as the mid-brain,

1239
00:54:02,040 --> 00:54:05,910
or what MacLean called
the mammalian brain.

1240
00:54:05,910 --> 00:54:09,660
This is the part of the
brain that we share only

1241
00:54:09,660 --> 00:54:12,630
with warm blooded
creatures, with mammals.

1242
00:54:12,630 --> 00:54:15,030
So reptiles don't have this.

1243
00:54:15,030 --> 00:54:18,480
And it's evolutionarily
somewhat later in development

1244
00:54:18,480 --> 00:54:20,010
than the reptilian brain.

1245
00:54:20,010 --> 00:54:22,290
And this part of the
brain is responsible

1246
00:54:22,290 --> 00:54:26,100
for emotional reaction,
for social behavior,

1247
00:54:26,100 --> 00:54:30,930
for fear and greed,
love, sexual attraction.

1248
00:54:30,930 --> 00:54:33,717
Now, you might ask, well, how
do neuroscientists know this?

1249
00:54:33,717 --> 00:54:35,550
Well, the way that
they've discovered this--

1250
00:54:35,550 --> 00:54:37,860
and it's only been over the
last couple of decades--

1251
00:54:37,860 --> 00:54:40,770
is they take a subject, and
they put him in an MRI machine.

1252
00:54:40,770 --> 00:54:42,690
All of you I think know
what an MRI is, right?

1253
00:54:42,690 --> 00:54:45,060
Magnetic resonance imaging.

1254
00:54:45,060 --> 00:54:48,300
And instead of taking a
snapshot, what neuroscientists

1255
00:54:48,300 --> 00:54:51,300
will do is to take a
series of snapshots

1256
00:54:51,300 --> 00:54:56,100
over many, many milliseconds
and eventually put together

1257
00:54:56,100 --> 00:54:58,120
a movie of your brain.

1258
00:54:58,120 --> 00:54:59,670
Now, the reason
that's interesting

1259
00:54:59,670 --> 00:55:05,970
is because blood contains iron,
and iron shows up on an MRI.

1260
00:55:05,970 --> 00:55:07,980
So when they take a
movie of your brain,

1261
00:55:07,980 --> 00:55:10,890
they see the blood
flowing in your brain.

1262
00:55:10,890 --> 00:55:12,960
And while a subject
is in there, they

1263
00:55:12,960 --> 00:55:17,550
will impose certain kinds
of stimuli on that subject

1264
00:55:17,550 --> 00:55:19,320
and see where the blood goes.

1265
00:55:19,320 --> 00:55:22,470
So, for example, they ask
you to lie in the magnet,

1266
00:55:22,470 --> 00:55:24,510
and you're lying
there, and they are

1267
00:55:24,510 --> 00:55:26,010
taking a movie of your brain.

1268
00:55:26,010 --> 00:55:27,840
And while you're lying
there peacefully,

1269
00:55:27,840 --> 00:55:30,180
they'll take a needle, and
they'll stab you with it.

1270
00:55:30,180 --> 00:55:33,180
And you go, "Ow!"

1271
00:55:33,180 --> 00:55:37,590
And what they'll show is
that blood flows to one very

1272
00:55:37,590 --> 00:55:40,400
specific region in the brain.

1273
00:55:40,400 --> 00:55:45,270
That region is the pain center
of the brain, or at least

1274
00:55:45,270 --> 00:55:46,740
that's how they identify it.

1275
00:55:46,740 --> 00:55:48,977
So when you feel
pain, and they'll

1276
00:55:48,977 --> 00:55:51,060
stab you with lots of
different parts of the body,

1277
00:55:51,060 --> 00:55:53,850
and they'll show that it
all leads to activation

1278
00:55:53,850 --> 00:55:55,950
of that little pain center.

1279
00:55:55,950 --> 00:55:58,920
So they'll say, fine,
pain is identified

1280
00:55:58,920 --> 00:56:01,450
in that particular region.

1281
00:56:01,450 --> 00:56:04,320
Now, the third part of
the brain, the last part

1282
00:56:04,320 --> 00:56:06,660
that MacLean
identified, is actually

1283
00:56:06,660 --> 00:56:08,010
what we think of as the brain.

1284
00:56:08,010 --> 00:56:10,362
It's that gray matter,
the curly coils.

1285
00:56:10,362 --> 00:56:11,820
It turns out that
that's not solid,

1286
00:56:11,820 --> 00:56:17,770
but that's like a blanket that
sits on top of the mid-brain.

1287
00:56:17,770 --> 00:56:19,980
And that part of the
brain is actually

1288
00:56:19,980 --> 00:56:24,700
unique to homo sapiens
and certain great apes.

1289
00:56:24,700 --> 00:56:26,947
So this is called
the hominid brain,

1290
00:56:26,947 --> 00:56:28,530
and, from an
evolutionary perspective,

1291
00:56:28,530 --> 00:56:30,790
it is the newest
part of the brain,

1292
00:56:30,790 --> 00:56:33,630
so it's often called the
neocortex, the newest

1293
00:56:33,630 --> 00:56:35,820
part of that brain.

1294
00:56:35,820 --> 00:56:39,060
Using MRI techniques,
neuroscientists

1295
00:56:39,060 --> 00:56:42,210
have been able to deduce
that this part of the brain

1296
00:56:42,210 --> 00:56:47,140
is responsible for language
ability, mathematical ability,

1297
00:56:47,140 --> 00:56:48,870
and logical deliberation.

1298
00:56:48,870 --> 00:56:53,170
What we think of as the higher
thought functions of humans.

1299
00:56:53,170 --> 00:56:56,902
It resides in that neocortex.

1300
00:56:56,902 --> 00:56:58,360
Now, why am I
telling you all this?

1301
00:56:58,360 --> 00:57:00,850
Well, because it turns out
that these three components

1302
00:57:00,850 --> 00:57:05,540
of the brain, they end up
working in different ways.

1303
00:57:05,540 --> 00:57:07,450
In fact, they have
different priorities.

1304
00:57:07,450 --> 00:57:08,420
What do I mean by that?

1305
00:57:08,420 --> 00:57:11,500
Well, let me give
you an example.

1306
00:57:11,500 --> 00:57:14,130
Suppose, God forbid,
you're in a car accident.

1307
00:57:14,130 --> 00:57:17,940
You're bleeding, and your
body starts to shut down.

1308
00:57:17,940 --> 00:57:20,040
You go into shock.

1309
00:57:20,040 --> 00:57:25,300
And, one by one, your
organs begin to fail.

1310
00:57:25,300 --> 00:57:28,360
Of the three components
of the brain,

1311
00:57:28,360 --> 00:57:31,660
which do you think
will shut down last?

1312
00:57:31,660 --> 00:57:34,660
Which of these three
components will shut down last?

1313
00:57:34,660 --> 00:57:35,630
Yeah?

1314
00:57:35,630 --> 00:57:36,630
AUDIENCE: The reptilian?

1315
00:57:36,630 --> 00:57:37,080
ANDREW LO: That's right.

1316
00:57:37,080 --> 00:57:37,868
Why?

1317
00:57:37,868 --> 00:57:41,059
AUDIENCE: It's responsible
for your basic functions, so--

1318
00:57:41,059 --> 00:57:41,850
ANDREW LO: Exactly.

1319
00:57:41,850 --> 00:57:44,680
Once your reptilian brain shuts
down, you know, that's bye-bye.

1320
00:57:44,680 --> 00:57:45,500
That's the end.

1321
00:57:45,500 --> 00:57:46,350
Right?

1322
00:57:46,350 --> 00:57:50,160
So this explains why you
can lose consciousness

1323
00:57:50,160 --> 00:57:51,430
but still breathe.

1324
00:57:51,430 --> 00:57:54,660
It's because
consciousness is seated

1325
00:57:54,660 --> 00:57:58,470
in the hominid brain and certain
parts of the mammalian brain.

1326
00:57:58,470 --> 00:57:59,970
But you don't need
to be conscious

1327
00:57:59,970 --> 00:58:01,520
in order to be breathing.

1328
00:58:01,520 --> 00:58:03,550
Your reptilian
brain controls that.

1329
00:58:03,550 --> 00:58:07,860
So from a priority perspective,
the most important part

1330
00:58:07,860 --> 00:58:12,600
of your three components
is the reptilian brain.

1331
00:58:12,600 --> 00:58:15,420
Now let me ask you
the next question.

1332
00:58:15,420 --> 00:58:18,890
Before that, of
the remaining two,

1333
00:58:18,890 --> 00:58:21,280
the hominid brain and the
mammalian brain, which

1334
00:58:21,280 --> 00:58:22,520
do you think has priority?

1335
00:58:22,520 --> 00:58:27,397
Which shuts down later in the
case of a car accident where

1336
00:58:27,397 --> 00:58:28,230
you're bleeding out?

1337
00:58:31,130 --> 00:58:31,752
Yeah?

1338
00:58:31,752 --> 00:58:33,080
AUDIENCE: The mammalian?

1339
00:58:33,080 --> 00:58:33,680
ANDREW LO: That's right.

1340
00:58:33,680 --> 00:58:34,179
Why?

1341
00:58:34,179 --> 00:58:36,770
AUDIENCE: Because that's
where fight or flight is?

1342
00:58:36,770 --> 00:58:39,920
ANDREW LO: Fight or flight
is in the mammalian brain,

1343
00:58:39,920 --> 00:58:43,626
but why would you think
that that's higher priority?

1344
00:58:43,626 --> 00:58:45,562
AUDIENCE: 'Cause there's
certain scenarios

1345
00:58:45,562 --> 00:58:48,002
and you want to react
to those [INAUDIBLE]

1346
00:58:48,002 --> 00:58:48,710
ANDREW LO: Right.

1347
00:58:48,710 --> 00:58:50,930
From an evolutionary
perspective,

1348
00:58:50,930 --> 00:58:52,610
it's probably more
important for you

1349
00:58:52,610 --> 00:58:55,490
to be scared and
run like hell when

1350
00:58:55,490 --> 00:58:57,950
you're being confronted
by a saber tooth tiger

1351
00:58:57,950 --> 00:59:00,492
than for you to be able to
solve differential equations.

1352
00:59:00,492 --> 00:59:01,730
Right?

1353
00:59:01,730 --> 00:59:04,550
Even if you're from MIT.

1354
00:59:04,550 --> 00:59:09,140
And so as a result, from
a evolutionary standpoint

1355
00:59:09,140 --> 00:59:12,530
and, more importantly, from
a physiological standpoint,

1356
00:59:12,530 --> 00:59:17,120
the mammalian brain has
priority over the hominid brain.

1357
00:59:17,120 --> 00:59:19,640
Now, neuroscientists
have documented

1358
00:59:19,640 --> 00:59:23,900
that this kind of hardwiring
is even more significant

1359
00:59:23,900 --> 00:59:25,550
than we had thought.

1360
00:59:25,550 --> 00:59:26,960
Let me explain why.

1361
00:59:26,960 --> 00:59:29,070
They did the
following experiment.

1362
00:59:29,070 --> 00:59:30,080
Take a subject.

1363
00:59:30,080 --> 00:59:31,520
Put him in a magnet.

1364
00:59:31,520 --> 00:59:35,894
And they asked the subject
to solve arithmetic problems

1365
00:59:35,894 --> 00:59:37,060
while they're in the magnet.

1366
00:59:37,060 --> 00:59:39,185
Now, you can do that because,
if you're lying down,

1367
00:59:39,185 --> 00:59:43,010
they have a mirror where they
project the image of a computer

1368
00:59:43,010 --> 00:59:44,720
screen, and your
hands are free even

1369
00:59:44,720 --> 00:59:46,550
though your head and
your body are not.

1370
00:59:46,550 --> 00:59:48,425
And so, while you're
lying down in the magnet

1371
00:59:48,425 --> 00:59:49,966
and you're looking
up at this mirror,

1372
00:59:49,966 --> 00:59:51,070
you see a computer screen.

1373
00:59:51,070 --> 00:59:52,160
And with your hands--

1374
00:59:52,160 --> 00:59:54,560
They give you a mouse,
and you can move around

1375
00:59:54,560 --> 00:59:57,290
and solve arithmetic
prompts on the screen

1376
00:59:57,290 --> 00:59:59,060
while you're in the magnet.

1377
00:59:59,060 --> 01:00:02,780
And arithmetic is
seated at this--

1378
01:00:02,780 --> 01:00:03,860
the hominid brain level.

1379
01:00:03,860 --> 01:00:05,900
So they can see
the blood flowing

1380
01:00:05,900 --> 01:00:09,020
into that part of the brain
as you solve these arithmetic

1381
01:00:09,020 --> 01:00:10,810
problems.

1382
01:00:10,810 --> 01:00:15,370
While you're doing that, the
experimenter will take a needle

1383
01:00:15,370 --> 01:00:18,280
and stab you with it.

1384
01:00:18,280 --> 01:00:19,690
And you go, "Ow!"

1385
01:00:19,690 --> 01:00:22,420
And then it takes you longer
to solve those arithmetic

1386
01:00:22,420 --> 01:00:23,752
problems.

1387
01:00:23,752 --> 01:00:25,960
Now you're saying, this is
what they give grants for?

1388
01:00:25,960 --> 01:00:29,229
I mean, anybody can tell
you that when somebody's

1389
01:00:29,229 --> 01:00:31,270
trying to stab you, it's
going to take you longer

1390
01:00:31,270 --> 01:00:32,985
to solve arithmetic problems.

1391
01:00:32,985 --> 01:00:34,300
OK.

1392
01:00:34,300 --> 01:00:36,340
But here's the insight.

1393
01:00:36,340 --> 01:00:38,830
The insight is that
what they discovered

1394
01:00:38,830 --> 01:00:44,590
was that the flow of blood was
constricted to your neocortex

1395
01:00:44,590 --> 01:00:48,420
after that pain that
was implemented,

1396
01:00:48,420 --> 01:00:51,850
but it lasted for hours.

1397
01:00:51,850 --> 01:00:54,910
The restriction of blood
flow to your neocortex

1398
01:00:54,910 --> 01:00:58,150
lasted for hours
after that pinprick.

1399
01:00:58,150 --> 01:01:02,650
Not just seconds, not just
minutes, but hours afterward,

1400
01:01:02,650 --> 01:01:06,420
it took you longer to
solve arithmetic problems.

1401
01:01:06,420 --> 01:01:10,320
For all intents and
purposes, that pain stimulus

1402
01:01:10,320 --> 01:01:13,320
made you stupider
for a period of time.

1403
01:01:13,320 --> 01:01:17,610
You physiologically could
not solve those problems

1404
01:01:17,610 --> 01:01:20,640
as quickly as before.

1405
01:01:20,640 --> 01:01:24,840
And this is a stunning
result. Because what it says

1406
01:01:24,840 --> 01:01:29,430
is that there are periods
where, for purely physiological

1407
01:01:29,430 --> 01:01:33,480
reasons, we are going to
be inhibited from using

1408
01:01:33,480 --> 01:01:35,110
that part of the brain.

1409
01:01:35,110 --> 01:01:36,930
I'm going to give you
some examples of this

1410
01:01:36,930 --> 01:01:38,850
in your own behavior,
just to prove

1411
01:01:38,850 --> 01:01:41,310
to you that this is
actually going on right now

1412
01:01:41,310 --> 01:01:43,450
in your heads.

1413
01:01:43,450 --> 01:01:46,000
How many of you have had
the following experience?

1414
01:01:46,000 --> 01:01:49,990
You're trying to meet a
very attractive individual.

1415
01:01:49,990 --> 01:01:51,850
You want to go out
on a date with them.

1416
01:01:51,850 --> 01:01:56,380
So you think of all sorts
of really clever ways

1417
01:01:56,380 --> 01:01:59,050
of meeting them by accident.

1418
01:01:59,050 --> 01:02:00,807
And then, when you
do that, you're

1419
01:02:00,807 --> 01:02:02,140
going to ask them out on a date.

1420
01:02:02,140 --> 01:02:03,730
And you've got your
lines prepared.

1421
01:02:03,730 --> 01:02:04,600
You know exactly
what you're going

1422
01:02:04,600 --> 01:02:06,460
to say to make them
just completely

1423
01:02:06,460 --> 01:02:09,000
thrilled at going out with you.

1424
01:02:09,000 --> 01:02:10,740
And finally, it
starts to unfold.

1425
01:02:10,740 --> 01:02:11,850
It happens.

1426
01:02:11,850 --> 01:02:15,510
You meet the person and, as
soon as you open your mouth,

1427
01:02:15,510 --> 01:02:18,450
and you start talking, you
sound like a total idiot.

1428
01:02:18,450 --> 01:02:21,210
You say all the
wrong things, right?

1429
01:02:21,210 --> 01:02:23,640
You're tongue tied,
you start stuttering.

1430
01:02:23,640 --> 01:02:25,739
And there's a reason for this.

1431
01:02:25,739 --> 01:02:27,030
There's a physiological reason.

1432
01:02:27,030 --> 01:02:28,480
You know what it is?

1433
01:02:28,480 --> 01:02:32,790
It's the fact that sexual
attraction hyperstimulates

1434
01:02:32,790 --> 01:02:35,730
your mid-brain, here.

1435
01:02:35,730 --> 01:02:38,070
And when it does, it
restricts your blood flow

1436
01:02:38,070 --> 01:02:39,330
to the neocortex.

1437
01:02:39,330 --> 01:02:43,410
Language ability is
in the neocortex.

1438
01:02:43,410 --> 01:02:44,820
Love makes you stupid.

1439
01:02:48,510 --> 01:02:51,370
Let me give you another example.

1440
01:02:51,370 --> 01:02:54,330
This is an example that we're
going to do together, OK?

1441
01:02:54,330 --> 01:02:56,830
The example is going
to go like this.

1442
01:02:56,830 --> 01:03:00,120
I'm going to show
you a bunch of words

1443
01:03:00,120 --> 01:03:02,230
that have different colors.

1444
01:03:02,230 --> 01:03:04,290
And I don't want you
to read anything,

1445
01:03:04,290 --> 01:03:07,860
but I want you to say out loud
with me what the colors are,

1446
01:03:07,860 --> 01:03:09,240
from left to right.

1447
01:03:09,240 --> 01:03:10,050
OK?

1448
01:03:10,050 --> 01:03:11,890
So these are the words.

1449
01:03:11,890 --> 01:03:14,700
But we're going to say out
loud together, as quickly

1450
01:03:14,700 --> 01:03:17,340
as possible, the colors.

1451
01:03:17,340 --> 01:03:19,346
You guys ready?

1452
01:03:19,346 --> 01:03:19,970
OK, here we go.

1453
01:03:19,970 --> 01:03:23,570
Red, green, blue, yellow,
orange, blue, brown, red,

1454
01:03:23,570 --> 01:03:26,630
green, purple, pink,
black, blue, yellow, green.

1455
01:03:26,630 --> 01:03:27,500
No problem, right?

1456
01:03:27,500 --> 01:03:29,490
Very easy.

1457
01:03:29,490 --> 01:03:32,680
We do the exact same
thing now with this.

1458
01:03:32,680 --> 01:03:36,230
Say the colors as
quickly as possible.

1459
01:03:36,230 --> 01:03:37,140
You ready?

1460
01:03:37,140 --> 01:03:38,030
OK, here we go.

1461
01:03:38,030 --> 01:03:40,740
Blue, red, blue, black, orange--

1462
01:03:40,740 --> 01:03:44,540
[LAUGHTER]

1463
01:03:44,540 --> 01:03:46,440
What happened?

1464
01:03:46,440 --> 01:03:48,340
What's going on?

1465
01:03:48,340 --> 01:03:51,950
Couldn't do it as
quickly, could you?

1466
01:03:51,950 --> 01:03:54,980
Well, let me tell
you what's going on.

1467
01:03:54,980 --> 01:04:00,140
It turns out that language
ability and color recognition

1468
01:04:00,140 --> 01:04:02,600
are actually handled by two
different parts of the brain.

1469
01:04:02,600 --> 01:04:05,060
Language ability
resides strictly

1470
01:04:05,060 --> 01:04:08,600
in the neocortex,
the hominid brain.

1471
01:04:08,600 --> 01:04:11,150
But color recognition
is actually

1472
01:04:11,150 --> 01:04:13,010
part of the mammalian brain.

1473
01:04:13,010 --> 01:04:15,999
Mammals can recognize
colors very well.

1474
01:04:15,999 --> 01:04:17,540
In fact, I don't
know how many of you

1475
01:04:17,540 --> 01:04:21,710
have redecorated your homes, or
your kitchens, or dining rooms,

1476
01:04:21,710 --> 01:04:22,550
recently.

1477
01:04:22,550 --> 01:04:24,390
I did this a couple
of years ago.

1478
01:04:24,390 --> 01:04:26,270
And my wife talked
to one of her friends

1479
01:04:26,270 --> 01:04:28,430
who is a decorator
who told her, you've

1480
01:04:28,430 --> 01:04:31,550
got to paint your
dining room in red.

1481
01:04:31,550 --> 01:04:32,660
Because-- Why?

1482
01:04:32,660 --> 01:04:34,100
Anybody know?

1483
01:04:34,100 --> 01:04:37,230
Why is red the preferred color
of restaurants, dining rooms?

1484
01:04:37,230 --> 01:04:37,730
Yeah?

1485
01:04:37,730 --> 01:04:38,720
AUDIENCE: Makes you eat more.

1486
01:04:38,720 --> 01:04:39,230
ANDREW LO: Exactly.

1487
01:04:39,230 --> 01:04:40,021
Makes you eat more.

1488
01:04:40,021 --> 01:04:43,440
It's a more appetizing color.

1489
01:04:43,440 --> 01:04:46,012
Anybody ever wonder why that is?

1490
01:04:46,012 --> 01:04:47,970
Do you know why red is
a more appetizing color?

1491
01:04:47,970 --> 01:04:50,370
I hate to tell you because you
may not be appetized anymore.

1492
01:04:50,370 --> 01:04:51,360
AUDIENCE: Because of the color?

1493
01:04:51,360 --> 01:04:52,818
ANDREW LO: It's
the color of blood.

1494
01:04:52,818 --> 01:04:54,450
Because it's hardwired into us.

1495
01:04:54,450 --> 01:04:57,074
When we see blood,
it's dinnertime.

1496
01:04:57,074 --> 01:04:59,722
[LAUGHTER]

1497
01:04:59,722 --> 01:05:03,240
And that is hardwired.

1498
01:05:03,240 --> 01:05:04,294
It's true.

1499
01:05:04,294 --> 01:05:06,210
I mean, I don't know how
many of you have ever

1500
01:05:06,210 --> 01:05:07,500
gone to The Blue Room.

1501
01:05:07,500 --> 01:05:09,420
That's a restaurant
down the street.

1502
01:05:09,420 --> 01:05:11,010
I've actually never
been there, and I

1503
01:05:11,010 --> 01:05:13,470
have to confess the
reason is that the idea

1504
01:05:13,470 --> 01:05:16,950
of a blue room for dinner is
not very appetizing to me.

1505
01:05:16,950 --> 01:05:20,520
Blue is not a color you
want to see in your food.

1506
01:05:20,520 --> 01:05:23,190
So color recognition
is a deep-seated part

1507
01:05:23,190 --> 01:05:24,630
of our mammalian brain.

1508
01:05:24,630 --> 01:05:26,640
But language ability is not.

1509
01:05:26,640 --> 01:05:29,550
Language ability is a
different part of the brain.

1510
01:05:29,550 --> 01:05:34,290
In this case, where the colors
and the words don't match,

1511
01:05:34,290 --> 01:05:36,990
it takes your brain
a split second longer

1512
01:05:36,990 --> 01:05:39,510
to adjudicate that conflict.

1513
01:05:39,510 --> 01:05:41,790
But in this case,
when they're the same,

1514
01:05:41,790 --> 01:05:43,590
you can rattle it off like this.

1515
01:05:43,590 --> 01:05:44,720
[SNAPS FINGERS]

1516
01:05:44,720 --> 01:05:47,470
So let me bring this
now all back to finance,

1517
01:05:47,470 --> 01:05:49,502
as I told you I would.

1518
01:05:49,502 --> 01:05:50,710
What's the point of all this?

1519
01:05:50,710 --> 01:05:54,010
The point is that
what we take to be

1520
01:05:54,010 --> 01:05:57,400
human preferences
for decision-making

1521
01:05:57,400 --> 01:06:00,580
is not a mathematical,
immutable object.

1522
01:06:00,580 --> 01:06:03,370
But, in fact, preferences
are the outcome

1523
01:06:03,370 --> 01:06:06,460
of interactions between
the mammalian brain

1524
01:06:06,460 --> 01:06:08,260
and the hominid brain.

1525
01:06:08,260 --> 01:06:14,250
And what we think of as
basic decision-making

1526
01:06:14,250 --> 01:06:18,310
requires the proper
balance of the two.

1527
01:06:18,310 --> 01:06:24,150
So, in other words, rationality
is not logical deliberation.

1528
01:06:24,150 --> 01:06:26,760
The patient, Elliot, was
a clear counterexample.

1529
01:06:26,760 --> 01:06:31,055
Elliot was very logical,
but he was not rational.

1530
01:06:31,055 --> 01:06:35,730
Emotion is a critical
component of rationality.

1531
01:06:35,730 --> 01:06:39,290
But too much emotion,
and you get stupid,

1532
01:06:39,290 --> 01:06:43,160
as the case with falling
in love or in the case

1533
01:06:43,160 --> 01:06:45,920
of overwhelming
emotional stimulus

1534
01:06:45,920 --> 01:06:50,060
cutting off blood flow
to your neocortex.

1535
01:06:50,060 --> 01:06:54,331
This, by the way, will lead to
some very useful self-help kind

1536
01:06:54,331 --> 01:06:54,830
of advice.

1537
01:06:54,830 --> 01:06:56,630
Let me give you an example.

1538
01:06:56,630 --> 01:06:59,480
Those of you who've ever been
involved in corporate politics,

1539
01:06:59,480 --> 01:07:01,610
back office politics--

1540
01:07:01,610 --> 01:07:04,820
one of the things that
you should keep in mind

1541
01:07:04,820 --> 01:07:09,170
is that, when you are
emotionally hyper-stimulated,

1542
01:07:09,170 --> 01:07:11,476
you will not be able
to think straight.

1543
01:07:11,476 --> 01:07:13,850
How many of you've ever heard
the phrase, I was so angry,

1544
01:07:13,850 --> 01:07:15,509
I could barely speak?

1545
01:07:15,509 --> 01:07:16,550
You've heard that, right?

1546
01:07:16,550 --> 01:07:17,930
You've probably felt that.

1547
01:07:17,930 --> 01:07:21,050
Well, that's not
just a metaphor.

1548
01:07:21,050 --> 01:07:22,820
That's actual
physiologically true.

1549
01:07:22,820 --> 01:07:26,120
When you are so angry, it
will cut off the flow of blood

1550
01:07:26,120 --> 01:07:28,610
to your neocortex, and you
will not be able to speak.

1551
01:07:28,610 --> 01:07:31,850
You will not be able to express
yourself as articulately.

1552
01:07:31,850 --> 01:07:34,100
So the next time somebody
goes into your office,

1553
01:07:34,100 --> 01:07:36,770
or your cubicle,
or your company,

1554
01:07:36,770 --> 01:07:40,040
and says something to
you that makes you so mad

1555
01:07:40,040 --> 01:07:45,000
that you want to strangle
them, and you can't speak,

1556
01:07:45,000 --> 01:07:48,210
the best thing to
do is not to speak.

1557
01:07:48,210 --> 01:07:50,060
It's to say, "Thank
you very much.

1558
01:07:50,060 --> 01:07:51,340
That was very insightful.

1559
01:07:51,340 --> 01:07:53,680
Let me get back to
you in a week or so."

1560
01:07:53,680 --> 01:07:56,440
Because if you try to talk
when you are emotionally

1561
01:07:56,440 --> 01:07:58,510
hyper-stimulated,
you will undoubtedly

1562
01:07:58,510 --> 01:08:00,610
say stupid things,
things you will regret

1563
01:08:00,610 --> 01:08:03,220
for possibly years later.

1564
01:08:03,220 --> 01:08:07,120
And this is a direct outcome
of this neurophysiological

1565
01:08:07,120 --> 01:08:08,560
research.

1566
01:08:08,560 --> 01:08:12,070
Our preferences are the sum
total of these interactions.

1567
01:08:12,070 --> 01:08:14,740
Therefore, they are
not stable over time,

1568
01:08:14,740 --> 01:08:17,080
nor are they stable
over circumstances.

1569
01:08:17,080 --> 01:08:17,649
All right?

1570
01:08:17,649 --> 01:08:21,100
You are not the same person
that you were 10 years ago.

1571
01:08:21,100 --> 01:08:24,880
And many of you who are going
to be facing losses of one kind

1572
01:08:24,880 --> 01:08:28,149
or another-- you lose a parent,
you lose a spouse or child--

1573
01:08:28,149 --> 01:08:30,970
you will not be the same person
after that event as before.

1574
01:08:30,970 --> 01:08:33,250
It will change you
because it will

1575
01:08:33,250 --> 01:08:36,729
change the balance
between your hominid brain

1576
01:08:36,729 --> 01:08:38,290
and your mammalian brain.

1577
01:08:38,290 --> 01:08:40,750
And unless you bring
those back into balance,

1578
01:08:40,750 --> 01:08:43,250
you're going to be making
decisions differently.

1579
01:08:43,250 --> 01:08:45,100
So if there's any
advice that I have

1580
01:08:45,100 --> 01:08:47,640
for you about how to take
this information and use it,

1581
01:08:47,640 --> 01:08:50,710
it's to keep your
eyes on the objective.

1582
01:08:50,710 --> 01:08:56,430
In other words, decide what you
want to achieve and then ask

1583
01:08:56,430 --> 01:08:59,340
yourself at every
step of the way,

1584
01:08:59,340 --> 01:09:04,290
"Is my current actions
about to help or hinder

1585
01:09:04,290 --> 01:09:05,729
those objectives?"

1586
01:09:05,729 --> 01:09:08,609
You know, if you're driving
home from work one day,

1587
01:09:08,609 --> 01:09:10,770
and somebody cuts
you off, and you

1588
01:09:10,770 --> 01:09:13,319
want to give them the finger
and ram your car into them,

1589
01:09:13,319 --> 01:09:15,720
ask yourself, will
that help you achieve

1590
01:09:15,720 --> 01:09:20,250
your objective of getting home
without an incident or not?

1591
01:09:20,250 --> 01:09:22,439
If it doesn't, you
probably shouldn't do it,

1592
01:09:22,439 --> 01:09:25,080
even though it'd feel
really good to do that.

1593
01:09:25,080 --> 01:09:26,250
OK?

1594
01:09:26,250 --> 01:09:30,029
Now, let me tell you about the
adaptive markets hypothesis

1595
01:09:30,029 --> 01:09:33,040
and how I'm going to put
it all together for you.

1596
01:09:33,040 --> 01:09:36,189
Behavioral finance
and rational finance.

1597
01:09:36,189 --> 01:09:39,640
They are both correct, and
they're both incorrect.

1598
01:09:39,640 --> 01:09:41,529
The reason that
they're both correct

1599
01:09:41,529 --> 01:09:46,720
is that they both apply
to certain circumstances.

1600
01:09:46,720 --> 01:09:48,760
What I taught you in
this course up until now,

1601
01:09:48,760 --> 01:09:52,029
what we've learned together,
is the rational finance

1602
01:09:52,029 --> 01:09:54,760
that applies most
of the time when

1603
01:09:54,760 --> 01:09:56,980
your mid-brain
and your neocortex

1604
01:09:56,980 --> 01:09:58,750
are properly balanced.

1605
01:09:58,750 --> 01:10:01,570
But during extreme
periods of distress,

1606
01:10:01,570 --> 01:10:04,960
when you are emotionally
either overwhelmed positive

1607
01:10:04,960 --> 01:10:08,950
or overwhelmed negative, in
either of those circumstances,

1608
01:10:08,950 --> 01:10:13,660
you will not behave in ways
that we would consider rational.

1609
01:10:13,660 --> 01:10:17,080
Those are the moments when
behavioral finance applies

1610
01:10:17,080 --> 01:10:19,300
or, at least, when
the anomalies arise.

1611
01:10:19,300 --> 01:10:21,220
Behavioral finance
is not a theory,

1612
01:10:21,220 --> 01:10:24,040
it's simply a
collection of anomalies.

1613
01:10:24,040 --> 01:10:26,830
The adaptive markets
hypothesis is this theory

1614
01:10:26,830 --> 01:10:29,230
that I've been developing for
the last several years that

1615
01:10:29,230 --> 01:10:31,030
tries to put these two together.

1616
01:10:31,030 --> 01:10:33,070
And the way to do
that is by recognizing

1617
01:10:33,070 --> 01:10:36,370
that we are both creatures
of our mammalian brain

1618
01:10:36,370 --> 01:10:38,680
and creatures of our neocortex.

1619
01:10:38,680 --> 01:10:42,160
Depending on market conditions,
environmental conditions,

1620
01:10:42,160 --> 01:10:44,810
and our own
decision-making process,

1621
01:10:44,810 --> 01:10:48,730
we may or may not be
in the rational camp

1622
01:10:48,730 --> 01:10:52,000
or the behaviorial camp
at any point in time.

1623
01:10:52,000 --> 01:10:54,490
And when you look at the
market as a whole, an aggregate

1624
01:10:54,490 --> 01:10:56,380
over all of these
individuals, you

1625
01:10:56,380 --> 01:10:58,540
will be able to get a
sense of whether or not

1626
01:10:58,540 --> 01:11:00,670
you can trust market prices.

1627
01:11:00,670 --> 01:11:03,689
At the beginning of the course,
as this crisis was unfolding,

1628
01:11:03,689 --> 01:11:05,480
I told you very clearly
that finance theory

1629
01:11:05,480 --> 01:11:08,050
is going to be taking a
several week vacation.

1630
01:11:08,050 --> 01:11:11,200
What I was getting at is
that, right now, all of us

1631
01:11:11,200 --> 01:11:14,350
are being hyper-stimulated
in our mid-brain.

1632
01:11:14,350 --> 01:11:17,470
We are all scared to death as to
what we're going to read about.

1633
01:11:17,470 --> 01:11:19,930
There was a time during
this course where

1634
01:11:19,930 --> 01:11:22,106
I was really not looking
forward to the weekends

1635
01:11:22,106 --> 01:11:24,730
because something always seemed
to be going on on the weekends.

1636
01:11:24,730 --> 01:11:27,313
And Monday I'd have to tell you,
well, gee, this bank is gone,

1637
01:11:27,313 --> 01:11:29,140
and that investment
bank is no more.

1638
01:11:29,140 --> 01:11:31,720
And it was a very
stressful period.

1639
01:11:31,720 --> 01:11:33,450
It still is.

1640
01:11:33,450 --> 01:11:36,130
Decision-making is going
to be affected by that.

1641
01:11:36,130 --> 01:11:39,300
So we developed
heuristics, rules of thumb,

1642
01:11:39,300 --> 01:11:45,090
to balance out our neocortex
and our mammalian brain.

1643
01:11:45,090 --> 01:11:49,320
We come up with, not
theoretical optimum optimization

1644
01:11:49,320 --> 01:11:52,030
algorithms, but we come
up with rules of thumb.

1645
01:11:52,030 --> 01:11:55,560
And the idea behind the
adaptive markets hypothesis

1646
01:11:55,560 --> 01:11:59,460
is that these rules of thumb
evolve over time as they either

1647
01:11:59,460 --> 01:12:01,770
succeed or they fail.

1648
01:12:01,770 --> 01:12:05,430
So the economics that we
learn and teach you in class

1649
01:12:05,430 --> 01:12:09,390
is meant to be an approximation
to a much more complex reality.

1650
01:12:09,390 --> 01:12:11,820
Sometimes that
approximation works,

1651
01:12:11,820 --> 01:12:14,010
sometimes it doesn't,
and the only way

1652
01:12:14,010 --> 01:12:16,890
to really understand
which it is,

1653
01:12:16,890 --> 01:12:19,890
is to try to get a sense of
how this evolutionary process

1654
01:12:19,890 --> 01:12:20,760
works.

1655
01:12:20,760 --> 01:12:23,640
In particular, where do
these various different kinds

1656
01:12:23,640 --> 01:12:26,530
of rules of thumb come from?

1657
01:12:26,530 --> 01:12:30,550
And to do that, I want to
tell you about one last thing,

1658
01:12:30,550 --> 01:12:32,020
and then we're
going to conclude,

1659
01:12:32,020 --> 01:12:35,500
which is that evolutionary
forces are really

1660
01:12:35,500 --> 01:12:39,040
what drives the
success or failure

1661
01:12:39,040 --> 01:12:41,260
of these rules of thumb.

1662
01:12:41,260 --> 01:12:44,080
And I want to give you an
example of these kinds of rules

1663
01:12:44,080 --> 01:12:45,910
of thumb and how
adaptation works.

1664
01:12:45,910 --> 01:12:47,860
And to do that, I'm
going to tell you

1665
01:12:47,860 --> 01:12:52,510
about a personal example
that has to do with a problem

1666
01:12:52,510 --> 01:12:54,370
that I solve every day.

1667
01:12:54,370 --> 01:12:57,740
And it's a problem that
I think you have as well.

1668
01:12:57,740 --> 01:13:00,490
I'm going to tell you about
how my heuristic developed

1669
01:13:00,490 --> 01:13:02,050
over the years.

1670
01:13:02,050 --> 01:13:03,850
The problem that I'm
going to talk about

1671
01:13:03,850 --> 01:13:05,990
is the problem of getting
dressed in the morning.

1672
01:13:05,990 --> 01:13:06,490
OK?

1673
01:13:06,490 --> 01:13:08,860
Every morning, all of
us have to get dressed.

1674
01:13:08,860 --> 01:13:11,440
And to give you a
sense of the challenges

1675
01:13:11,440 --> 01:13:14,600
that I face in my process
of getting dressed,

1676
01:13:14,600 --> 01:13:16,360
I have to tell you
about my wardrobe.

1677
01:13:16,360 --> 01:13:18,340
So here's my wardrobe.

1678
01:13:18,340 --> 01:13:20,500
I've got five jackets,
10 pairs of pants,

1679
01:13:20,500 --> 01:13:23,290
20 ties, 10 shirts, 10 pairs
of socks, four pairs of shoes,

1680
01:13:23,290 --> 01:13:25,730
and five belts.

1681
01:13:25,730 --> 01:13:28,520
Now, you might argue that's
a rather limited wardrobe.

1682
01:13:28,520 --> 01:13:33,020
However, if you took the time
to compute the combinatorics,

1683
01:13:33,020 --> 01:13:36,800
you would see that I have
two million unique outfits

1684
01:13:36,800 --> 01:13:38,420
in my wardrobe.

1685
01:13:38,420 --> 01:13:42,530
Now, granted, not all of
them are of the same fashion

1686
01:13:42,530 --> 01:13:43,610
content.

1687
01:13:43,610 --> 01:13:45,560
They're not all
equally compelling

1688
01:13:45,560 --> 01:13:47,660
from the fashion
perspective, so I've

1689
01:13:47,660 --> 01:13:48,890
got an optimization problem.

1690
01:13:48,890 --> 01:13:51,290
I've got to figure out
which is the right outfit

1691
01:13:51,290 --> 01:13:53,040
for the right occasion.

1692
01:13:53,040 --> 01:13:56,390
Suppose it takes me one
second to evaluate the fashion

1693
01:13:56,390 --> 01:13:59,250
content of an outfit.

1694
01:13:59,250 --> 01:14:02,130
How long would it take me
to get dressed every day?

1695
01:14:02,130 --> 01:14:04,350
Well, you do the math.

1696
01:14:04,350 --> 01:14:07,674
Turns out that it would take
me 23.1 days to get dressed.

1697
01:14:07,674 --> 01:14:09,840
Now, clearly, I don't take
that long to get dressed,

1698
01:14:09,840 --> 01:14:11,250
so how do I do it?

1699
01:14:11,250 --> 01:14:14,370
Do I have some superhuman
optimization algorithm up

1700
01:14:14,370 --> 01:14:16,800
my sleeve?

1701
01:14:16,800 --> 01:14:17,820
No.

1702
01:14:17,820 --> 01:14:20,427
It turns out I use a
heuristic, a rule of thumb.

1703
01:14:20,427 --> 01:14:23,010
And you might say, well, where
did you get that rule of thumb?

1704
01:14:23,010 --> 01:14:24,930
I'll tell you.

1705
01:14:24,930 --> 01:14:27,600
When I was six years old,
growing up in Queens,

1706
01:14:27,600 --> 01:14:31,050
New York, at that time
the superhero of the day

1707
01:14:31,050 --> 01:14:32,820
was Superman.

1708
01:14:32,820 --> 01:14:36,747
Everybody watched Superman on
TV, comic books, and so on.

1709
01:14:36,747 --> 01:14:38,580
And some clever marketing
genius figured out

1710
01:14:38,580 --> 01:14:41,370
that if you put a Superman
emblem on a jacket,

1711
01:14:41,370 --> 01:14:44,520
you could sell a lot of
them to six-year-old kids.

1712
01:14:44,520 --> 01:14:48,070
And growing up in a
single parent household,

1713
01:14:48,070 --> 01:14:50,190
we didn't have a
lot of extra cash,

1714
01:14:50,190 --> 01:14:54,570
so it took me weeks to nag my
mother to get me this jacket.

1715
01:14:54,570 --> 01:14:56,850
Weeks and weeks
of daily nagging.

1716
01:14:56,850 --> 01:14:58,840
And finally, after
weeks of nagging,

1717
01:14:58,840 --> 01:15:02,560
she relented, and she agreed
to buy me this jacket.

1718
01:15:02,560 --> 01:15:03,960
And I remember
vividly, you know,

1719
01:15:03,960 --> 01:15:06,710
Friday afternoon after school,
she got home from work.

1720
01:15:06,710 --> 01:15:09,239
I waited for her, we went to
Alexander's, bought the jacket.

1721
01:15:09,239 --> 01:15:11,280
I spent the whole weekend
wearing this thing, day

1722
01:15:11,280 --> 01:15:11,780
and night.

1723
01:15:11,780 --> 01:15:13,500
It was wonderful.

1724
01:15:13,500 --> 01:15:17,556
And Monday morning, I got up
extra early, really excited

1725
01:15:17,556 --> 01:15:18,930
to go to school
with this jacket.

1726
01:15:18,930 --> 01:15:20,888
I looked in the mirror,
did all my action poses

1727
01:15:20,888 --> 01:15:23,220
and looked at myself.

1728
01:15:23,220 --> 01:15:26,130
And by the time I
was done with that,

1729
01:15:26,130 --> 01:15:28,710
I was 15 minutes
late for school,

1730
01:15:28,710 --> 01:15:32,650
and I needed a note from my
mother to get into class.

1731
01:15:32,650 --> 01:15:35,490
And I remember vividly,
walking into the classroom,

1732
01:15:35,490 --> 01:15:37,620
going up to the
front of the room,

1733
01:15:37,620 --> 01:15:40,110
giving the teacher
the note, walking back

1734
01:15:40,110 --> 01:15:42,450
to my seat, all the other
kids already sitting down,

1735
01:15:42,450 --> 01:15:44,520
reading, snickering at me.

1736
01:15:44,520 --> 01:15:47,010
And I was completely mortified.

1737
01:15:47,010 --> 01:15:49,650
And you know I must be mortified
because, 42 years later, I

1738
01:15:49,650 --> 01:15:52,710
still remember this exactly.

1739
01:15:52,710 --> 01:15:53,700
And you know what?

1740
01:15:53,700 --> 01:15:57,900
From that day on, it never
took me more than five minutes

1741
01:15:57,900 --> 01:15:59,130
to get dressed.

1742
01:15:59,130 --> 01:15:59,970
Ever.

1743
01:15:59,970 --> 01:16:02,250
Never.

1744
01:16:02,250 --> 01:16:07,650
So I solved my problem
based upon very strong,

1745
01:16:07,650 --> 01:16:10,020
negative feedback.

1746
01:16:10,020 --> 01:16:11,274
You see how it works?

1747
01:16:11,274 --> 01:16:13,690
I have a new heuristic for
getting dressed in the morning,

1748
01:16:13,690 --> 01:16:17,100
but that new heuristic only
came about after some very

1749
01:16:17,100 --> 01:16:20,270
strong, negative emotion.

1750
01:16:20,270 --> 01:16:22,490
If I was like
Elliot, and I never

1751
01:16:22,490 --> 01:16:25,800
had any negative
feedback, I'd go

1752
01:16:25,800 --> 01:16:29,070
on taking hours to get
dressed every morning

1753
01:16:29,070 --> 01:16:32,190
instead of getting dressed in
five minutes, like I do now.

1754
01:16:32,190 --> 01:16:36,429
Now, somebody else, let's say
a movie star like Tom Cruise.

1755
01:16:36,429 --> 01:16:38,970
I suspect that Tom Cruise spends
a lot more than five minutes

1756
01:16:38,970 --> 01:16:40,620
getting dressed in the morning.

1757
01:16:40,620 --> 01:16:43,110
In fact, he probably spends
more time coiffing his hair

1758
01:16:43,110 --> 01:16:46,310
than I spend coiffing my hair.

1759
01:16:46,310 --> 01:16:48,360
And there's a reason for that.

1760
01:16:48,360 --> 01:16:50,790
For him, it makes sense.

1761
01:16:50,790 --> 01:16:54,510
That heuristic for him
makes sense in his context,

1762
01:16:54,510 --> 01:16:56,760
not in my context.

1763
01:16:56,760 --> 01:17:00,620
All of us, we engage
in these heuristics

1764
01:17:00,620 --> 01:17:04,890
that are adaptive to the
particular situation at hand.

1765
01:17:04,890 --> 01:17:06,570
So the adaptive
markets hypothesis

1766
01:17:06,570 --> 01:17:08,610
acknowledges that
people make mistakes,

1767
01:17:08,610 --> 01:17:11,610
but they learn, and
they adapt, assuming

1768
01:17:11,610 --> 01:17:16,410
there's balance between emotion
and logical deliberation.

1769
01:17:16,410 --> 01:17:20,400
So there are a number of
very practical implications

1770
01:17:20,400 --> 01:17:25,460
for this, some of which
suggest that the CAPM does not

1771
01:17:25,460 --> 01:17:29,540
work and should not work,
except during certain periods

1772
01:17:29,540 --> 01:17:30,870
of the market.

1773
01:17:30,870 --> 01:17:34,940
So what we've learned
together in this course

1774
01:17:34,940 --> 01:17:40,040
describes 80% to 90% of
what happens in markets.

1775
01:17:40,040 --> 01:17:44,450
But there's another 10% or 20%
of extreme and unusual market

1776
01:17:44,450 --> 01:17:48,260
conditions that makes this field
a lot more challenging than you

1777
01:17:48,260 --> 01:17:50,060
might otherwise think.

1778
01:17:50,060 --> 01:17:52,490
And that's the part that I
want you to keep in mind.

1779
01:17:52,490 --> 01:17:55,520
We're not going to talk about
that anymore in this course,

1780
01:17:55,520 --> 01:17:59,480
or in 402, or in 433,
or in 434, or 437.

1781
01:17:59,480 --> 01:18:02,210
We don't talk about the
messy part of finance

1782
01:18:02,210 --> 01:18:04,230
because it's still
currently being developed.

1783
01:18:04,230 --> 01:18:05,810
In fact, as I told you
at the very beginning,

1784
01:18:05,810 --> 01:18:07,351
you're not going to
find any textbook

1785
01:18:07,351 --> 01:18:09,680
versions of the adaptive
markets hypothesis.

1786
01:18:09,680 --> 01:18:11,721
The only person you'll
ever find talking about it

1787
01:18:11,721 --> 01:18:14,330
right now is me, unfortunately.

1788
01:18:14,330 --> 01:18:15,990
Hopefully, that will
change over time,

1789
01:18:15,990 --> 01:18:17,630
but this is still
under development.

1790
01:18:17,630 --> 01:18:21,410
But just keep in mind that the
theories that you've learned

1791
01:18:21,410 --> 01:18:24,620
apply not all of the time.

1792
01:18:24,620 --> 01:18:27,590
Most of the time, they
apply and they're useful,

1793
01:18:27,590 --> 01:18:31,040
but they are an approximation
to a much more complex reality.

1794
01:18:31,040 --> 01:18:33,590
And that reality
will incorporate

1795
01:18:33,590 --> 01:18:37,220
lots of other phenomenon
which we described here

1796
01:18:37,220 --> 01:18:41,020
and which I'll talk a
bit more about on Monday.

1797
01:18:41,020 --> 01:18:41,970
OK.

1798
01:18:41,970 --> 01:18:44,620
So let me stop here since
we're just about out of time.

1799
01:18:44,620 --> 01:18:47,151
Any questions?

1800
01:18:47,151 --> 01:18:47,650
No?

1801
01:18:47,650 --> 01:18:48,380
Yes?

1802
01:18:48,380 --> 01:18:52,980
AUDIENCE: So I was wondering if
you might think, for example,

1803
01:18:52,980 --> 01:18:56,500
if I have an exam,
I tend to notice

1804
01:18:56,500 --> 01:19:00,560
that I will be way more
productive at learning way

1805
01:19:00,560 --> 01:19:03,508
before two days before.

1806
01:19:03,508 --> 01:19:07,950
And I cannot see a conflict
because I was feeling

1807
01:19:07,950 --> 01:19:10,100
the pressure in
handling my emotions.

1808
01:19:10,100 --> 01:19:12,860
But how is it possible that
I feel the emotion but,

1809
01:19:12,860 --> 01:19:14,644
at the same time, I--

1810
01:19:14,644 --> 01:19:16,310
ANDREW LO: That is a
fantastic question.

1811
01:19:16,310 --> 01:19:17,240
Let me repeat the question.

1812
01:19:17,240 --> 01:19:18,823
I'm not going to
answer today, but I'm

1813
01:19:18,823 --> 01:19:20,780
going to talk about it
on Monday, because it's

1814
01:19:20,780 --> 01:19:24,710
both a great question in
the context of this theory,

1815
01:19:24,710 --> 01:19:26,370
but it's also very
apropos, so I'm

1816
01:19:26,370 --> 01:19:28,760
going to be talking about
the final exam next Monday.

1817
01:19:28,760 --> 01:19:29,510
OK?

1818
01:19:29,510 --> 01:19:32,134
The question is, when you
were studying for an exam,

1819
01:19:32,134 --> 01:19:33,050
you're under pressure.

1820
01:19:33,050 --> 01:19:35,150
You're feeling a lot
of emotional stress,

1821
01:19:35,150 --> 01:19:39,050
and yet you seem to
learn better, faster,

1822
01:19:39,050 --> 01:19:41,580
during those periods of time.

1823
01:19:41,580 --> 01:19:45,410
It turns out that there is a
physiological and psychological

1824
01:19:45,410 --> 01:19:48,530
basis when a certain
amount of pressure

1825
01:19:48,530 --> 01:19:50,149
is actually a good thing.

1826
01:19:50,149 --> 01:19:52,190
But the question is, what
is that certain amount?

1827
01:19:52,190 --> 01:19:55,367
Any more than that, and
then you'd be a basket case.

1828
01:19:55,367 --> 01:19:57,200
So we're going to talk
about that on Monday.

1829
01:19:57,200 --> 01:19:57,700
All right?

1830
01:19:57,700 --> 01:19:59,110
Thanks.