1
00:00:00,000 --> 00:00:02,400
[SQUEAKING]

2
00:00:02,400 --> 00:00:04,800
[RUSTLING]

3
00:00:04,800 --> 00:00:10,275
[CLICKING]

4
00:00:10,275 --> 00:00:11,400
FRANK SCHILBACH: All right.

5
00:00:11,400 --> 00:00:12,442
I'm going to get started.

6
00:00:12,442 --> 00:00:16,650
This is Lecture 23, about
policy with behavioral agents,

7
00:00:16,650 --> 00:00:21,400
the last lecture of 14.13,
for better or worse.

8
00:00:21,400 --> 00:00:21,900
OK.

9
00:00:21,900 --> 00:00:23,108
So what's the plan for today?

10
00:00:23,108 --> 00:00:25,410
We're going to talk
about paternalism,

11
00:00:25,410 --> 00:00:27,890
what is paternalism,
reasons for paternalism,

12
00:00:27,890 --> 00:00:30,990
why it might be popular or not.

13
00:00:30,990 --> 00:00:34,320
We're going to talk about
the specific example of what

14
00:00:34,320 --> 00:00:37,155
was coined as
libertarian paternalism.

15
00:00:37,155 --> 00:00:39,360
I'll explain more, which
is the Save More Tomorrow

16
00:00:39,360 --> 00:00:42,150
plan, what's good about
this specific example.

17
00:00:42,150 --> 00:00:43,620
And we'll discuss that.

18
00:00:43,620 --> 00:00:47,050
And finally, we'll talk
about some market solutions.

19
00:00:47,050 --> 00:00:49,780
Often one solution
that's often mentioned

20
00:00:49,780 --> 00:00:52,120
is to say, well,
why can't the market

21
00:00:52,120 --> 00:00:53,320
solve these kinds of issues.

22
00:00:53,320 --> 00:00:55,750
And we're going to talk
about why that may or may not

23
00:00:55,750 --> 00:00:58,480
be a good idea.

24
00:00:58,480 --> 00:01:01,690
So then, let me
just ask very simply

25
00:01:01,690 --> 00:01:05,319
to start with, from
your classes that you've

26
00:01:05,319 --> 00:01:07,330
taken in economics
and otherwise,

27
00:01:07,330 --> 00:01:11,230
what are some reasons
for government policies?

28
00:01:11,230 --> 00:01:13,120
And it doesn't mean
that you necessarily

29
00:01:13,120 --> 00:01:14,050
endorse those reasons.

30
00:01:14,050 --> 00:01:16,050
But what are the reasons
why the government, why

31
00:01:16,050 --> 00:01:18,040
we might want the
government to engage

32
00:01:18,040 --> 00:01:22,640
in an economic or other sort
of of government policies?

33
00:01:22,640 --> 00:01:25,390
So one reason might be
that people don't quite

34
00:01:25,390 --> 00:01:27,420
know what's good for them.

35
00:01:27,420 --> 00:01:32,230
And so we can potentially
help them by doing so.

36
00:01:32,230 --> 00:01:34,450
That's, of course, the reason--

37
00:01:34,450 --> 00:01:37,700
that's, of course, what we're
going to talk about today.

38
00:01:37,700 --> 00:01:40,820
Let me rephrase my question.

39
00:01:40,820 --> 00:01:44,120
In neoclassical economics, what
are reasons for the government

40
00:01:44,120 --> 00:01:45,560
to engage?

41
00:01:45,560 --> 00:01:47,457
The behavioral ones
of course, we're

42
00:01:47,457 --> 00:01:49,790
going to call internalities,
which is essentially people

43
00:01:49,790 --> 00:01:51,830
must optimize in certain ways.

44
00:01:51,830 --> 00:01:54,420
But what are the
neoclassical reasons

45
00:01:54,420 --> 00:01:56,840
for the government to engage?

46
00:01:56,840 --> 00:02:00,120
So this would be things
like externalities.

47
00:02:00,120 --> 00:02:03,170
This could be individuals
causing externalities.

48
00:02:03,170 --> 00:02:05,240
Externalities, essentially,
is a fancy word

49
00:02:05,240 --> 00:02:10,430
for people causing some positive
or negative effects on others

50
00:02:10,430 --> 00:02:13,300
that they might not be
able to internalize.

51
00:02:13,300 --> 00:02:16,320
This could be smokers or
individuals doing that.

52
00:02:16,320 --> 00:02:21,270
Like, for example, smoking might
cause health effects on others.

53
00:02:21,270 --> 00:02:23,390
Could be also farms
polluting the environment

54
00:02:23,390 --> 00:02:28,280
or polluting water or some
other resources of other people

55
00:02:28,280 --> 00:02:30,208
polluting like--

56
00:02:30,208 --> 00:02:32,000
otherwise the air, and
so on, and so forth,

57
00:02:32,000 --> 00:02:33,980
which might be bad
for others' health.

58
00:02:33,980 --> 00:02:36,470
And to the extent that
the firm or others are not

59
00:02:36,470 --> 00:02:38,600
sort of taking
that into account,

60
00:02:38,600 --> 00:02:41,600
that will lead to socially
sub optimal outcomes.

61
00:02:41,600 --> 00:02:44,240
Now the government would
say, well, let's tax that.

62
00:02:44,240 --> 00:02:46,010
Because we want to
make sure that people

63
00:02:46,010 --> 00:02:49,970
are taking into account
these externalities.

64
00:02:49,970 --> 00:02:51,710
And with the optimal
tax schedule,

65
00:02:51,710 --> 00:02:56,750
potentially, these can get back
to the [INAUDIBLE] potentially.

66
00:02:56,750 --> 00:02:59,750
So usually the terms that look
like free markets, et cetera,

67
00:02:59,750 --> 00:03:01,760
they essentially
will-- the free market

68
00:03:01,760 --> 00:03:06,110
will achieve that things are
efficient and potentially

69
00:03:06,110 --> 00:03:07,980
under some conditions.

70
00:03:07,980 --> 00:03:10,880
But what the market will
not typically achieve

71
00:03:10,880 --> 00:03:13,040
is some form of equity.

72
00:03:13,040 --> 00:03:17,420
And to the extent that we
are interested in that,

73
00:03:17,420 --> 00:03:20,600
you might want to engage in
social security, insurance

74
00:03:20,600 --> 00:03:23,640
program, unemployment insurance,
and so on, and so forth.

75
00:03:23,640 --> 00:03:25,430
So if you have lots
of competition,

76
00:03:25,430 --> 00:03:28,610
the prices will fall down
to often marginal costs.

77
00:03:28,610 --> 00:03:31,130
And that's good for consumers
because stuff is cheap.

78
00:03:31,130 --> 00:03:32,690
It could [INAUDIBLE]
only one firm,

79
00:03:32,690 --> 00:03:35,660
the firm, the monopolist
in the [INAUDIBLE] case,

80
00:03:35,660 --> 00:03:39,390
the monopolist will charge
above a marginal cost.

81
00:03:39,390 --> 00:03:41,780
So prices will be higher,
fewer things will be sold,

82
00:03:41,780 --> 00:03:44,010
and people will be
less happy about that.

83
00:03:44,010 --> 00:03:48,080
And then there's sort of this
antitrust in 1420, as you say,

84
00:03:48,080 --> 00:03:51,170
exactly Nancy Rose and
others are working on

85
00:03:51,170 --> 00:03:54,410
that, antitrust
commissions, et cetera,

86
00:03:54,410 --> 00:03:56,660
trying to essentially
make sure there's

87
00:03:56,660 --> 00:03:58,370
sufficient
computation in markets

88
00:03:58,370 --> 00:04:01,190
because we think that's
good for welfare.

89
00:04:01,190 --> 00:04:04,190
So there's other sort
of market failures.

90
00:04:04,190 --> 00:04:07,270
For example, public
goods, parts, and so on,

91
00:04:07,270 --> 00:04:09,020
the government might
want to provide that.

92
00:04:09,020 --> 00:04:13,310
Because in general, people
might under-provide that.

93
00:04:13,310 --> 00:04:15,440
Could be also other
things, like information,

94
00:04:15,440 --> 00:04:19,640
[INAUDIBLE] information, tend
to fail, reason being that,

95
00:04:19,640 --> 00:04:21,930
when I come up
with an idea that's

96
00:04:21,930 --> 00:04:23,900
useful for a lot of
people, I might not

97
00:04:23,900 --> 00:04:26,510
be able to internalize
the rents from that

98
00:04:26,510 --> 00:04:30,897
if I'm under investing,
therefore, in R&D, and so on.

99
00:04:30,897 --> 00:04:33,230
So now I think we have all
the things I have on my list.

100
00:04:33,230 --> 00:04:35,480
There's macroeconomic policy,
as Tom just mentioned.

101
00:04:35,480 --> 00:04:38,810
The Federal Reserve might try
to [INAUDIBLE] its interest rate

102
00:04:38,810 --> 00:04:40,460
and fiscal stimulus.

103
00:04:40,460 --> 00:04:44,135
Competition policy, that's
what Even just mentioned.

104
00:04:44,135 --> 00:04:45,510
Then there's just
redistribution,

105
00:04:45,510 --> 00:04:46,350
social insurance.

106
00:04:46,350 --> 00:04:49,050
I think that's what Justin
said on unemployment insurance,

107
00:04:49,050 --> 00:04:50,680
social security, and so on.

108
00:04:50,680 --> 00:04:52,350
And then there might
be externality.

109
00:04:52,350 --> 00:04:55,350
That's what Natalie
said, particular taxes

110
00:04:55,350 --> 00:04:58,780
and other sort of market
failures, such as innovation.

111
00:04:58,780 --> 00:05:02,790
That's what Jose, as well, said.

112
00:05:02,790 --> 00:05:05,463
What we're going to focus on now
is none of these things which,

113
00:05:05,463 --> 00:05:06,880
and that's getting
back to what we

114
00:05:06,880 --> 00:05:09,870
said at the very beginning,
which is the internalities.

115
00:05:09,870 --> 00:05:13,882
So that's not a traditional
way for governments to engage.

116
00:05:13,882 --> 00:05:16,465
And if you went back, like, 20
years from now, the government,

117
00:05:16,465 --> 00:05:19,890
there would be no such urge
to use such considerations.

118
00:05:19,890 --> 00:05:23,432
But more recently,
in part, perhaps,

119
00:05:23,432 --> 00:05:25,640
as a consequence of the rise
of behavioral economics,

120
00:05:25,640 --> 00:05:28,860
people are much more interested
now in internalities.

121
00:05:28,860 --> 00:05:30,440
So what are internalities?

122
00:05:30,440 --> 00:05:33,430
It's essentially sort of
similar to [INAUDIBLE]..

123
00:05:33,430 --> 00:05:35,730
The coin is very similar.

124
00:05:35,730 --> 00:05:41,540
Instead of causing damage or
not internalizing sufficiently

125
00:05:41,540 --> 00:05:43,490
effects on other
people, this is now

126
00:05:43,490 --> 00:05:46,130
effects on your future
self, essentially.

127
00:05:46,130 --> 00:05:48,860
So that the consumer is not
fully [INAUDIBLE] the costs

128
00:05:48,860 --> 00:05:54,560
and/or benefits she imposes
on her future potential self.

129
00:05:54,560 --> 00:05:56,060
And if the government
is sort of now

130
00:05:56,060 --> 00:05:58,477
confident that that's the case,
well, then, the government

131
00:05:58,477 --> 00:06:00,810
might want to intervene.

132
00:06:00,810 --> 00:06:01,910
Now what's paternalism?

133
00:06:01,910 --> 00:06:05,870
Paternalism is essentially
a version of that.

134
00:06:05,870 --> 00:06:09,200
There's a very
convoluted definition

135
00:06:09,200 --> 00:06:11,952
by the Merriam-Webster
dictionary, which

136
00:06:11,952 --> 00:06:13,160
I'm not going to read to you.

137
00:06:13,160 --> 00:06:15,000
But it is one definition.

138
00:06:15,000 --> 00:06:17,930
And there's a better one
from David Laibson, which

139
00:06:17,930 --> 00:06:20,510
is "An attempt to influence
or control people's

140
00:06:20,510 --> 00:06:22,407
conduct for their own good.

141
00:06:22,407 --> 00:06:24,740
In other words, when the
motivation for the intervention

142
00:06:24,740 --> 00:06:26,240
is not about externalities."

143
00:06:26,240 --> 00:06:28,160
So it's to say if
we try to affect

144
00:06:28,160 --> 00:06:29,840
people's behavior
for their own good

145
00:06:29,840 --> 00:06:33,140
in some way, that sort of
assumes, of course, that people

146
00:06:33,140 --> 00:06:36,440
are not fully optimizing in
certain ways, not when it comes

147
00:06:36,440 --> 00:06:40,640
to their effects on
others, [INAUDIBLE]

148
00:06:40,640 --> 00:06:43,633
the effects on
themselves, [INAUDIBLE]..

149
00:06:43,633 --> 00:06:44,860
OK?

150
00:06:44,860 --> 00:06:45,360
[INAUDIBLE]

151
00:06:45,360 --> 00:06:46,423
There was a question?

152
00:06:50,770 --> 00:06:55,340
Now when is
paternalism warranted?

153
00:06:55,340 --> 00:06:58,430
Because [INAUDIBLE] going
through the list of topics

154
00:06:58,430 --> 00:07:00,740
that we discussed in the class.

155
00:07:00,740 --> 00:07:04,460
And it's important to understand
that not all of these topics

156
00:07:04,460 --> 00:07:07,520
are, in fact,
warranting intervention.

157
00:07:07,520 --> 00:07:10,410
For example, if you think people
have certain risk preferences,

158
00:07:10,410 --> 00:07:13,130
reference-dependent preferences,
or social preferences,

159
00:07:13,130 --> 00:07:15,470
that's not a bias or
problem that we need to fix.

160
00:07:15,470 --> 00:07:16,580
People care about others.

161
00:07:16,580 --> 00:07:19,250
And that's just the way their
utility function is like.

162
00:07:19,250 --> 00:07:26,540
And then you show that people
can optimize accordingly.

163
00:07:26,540 --> 00:07:29,643
That's not necessarily
a bias or a problem.

164
00:07:29,643 --> 00:07:32,060
Of course, if there's problems
with self-control, problems

165
00:07:32,060 --> 00:07:34,227
with people at [INAUDIBLE]
inconsistent preferences,

166
00:07:34,227 --> 00:07:35,750
we might want to help people.

167
00:07:35,750 --> 00:07:37,790
That assumes, of course,
that we know which self

168
00:07:37,790 --> 00:07:40,910
is the one we want to support.

169
00:07:40,910 --> 00:07:43,370
There's lots of
space here, when we

170
00:07:43,370 --> 00:07:45,020
think about
non-standard beliefs,

171
00:07:45,020 --> 00:07:48,980
for people to make mistakes, or
for our governments and others

172
00:07:48,980 --> 00:07:49,860
to intervene.

173
00:07:49,860 --> 00:07:52,410
This would be limited attention,
also things like memory,

174
00:07:52,410 --> 00:07:54,950
like, people just forget stuff.

175
00:07:54,950 --> 00:07:57,860
Learning failure is where people
have just wrong information.

176
00:07:57,860 --> 00:08:01,490
Overoptimism, overconfidence,
projection, and attribution

177
00:08:01,490 --> 00:08:03,590
bias, these are all
essentially people that

178
00:08:03,590 --> 00:08:04,940
make mistakes in some ways.

179
00:08:04,940 --> 00:08:07,640
And to the extent that
they make mistakes,

180
00:08:07,640 --> 00:08:09,360
we might want to intervene.

181
00:08:09,360 --> 00:08:11,210
Importantly, we
discussed this already,

182
00:08:11,210 --> 00:08:14,480
motivated beliefs are
not necessarily a reason,

183
00:08:14,480 --> 00:08:17,150
or going to be very
careful with that.

184
00:08:17,150 --> 00:08:20,660
That is, as we said previously
when we talked about health

185
00:08:20,660 --> 00:08:21,720
issues, for example.

186
00:08:21,720 --> 00:08:25,590
So if somebody does not
want to know their health

187
00:08:25,590 --> 00:08:28,980
status because they want to feel
happy or make themselves think

188
00:08:28,980 --> 00:08:32,492
that they are happy, and
they don't want to know that,

189
00:08:32,492 --> 00:08:34,659
because they want to have
a few years of their lives

190
00:08:34,659 --> 00:08:36,576
in which they think can
[INAUDIBLE] themselves

191
00:08:36,576 --> 00:08:40,500
into thinking they're
happy, healthy and happy,

192
00:08:40,500 --> 00:08:42,659
it's not obvious that
they're making a mistake.

193
00:08:42,659 --> 00:08:44,250
In a sense, they could
be making a mistake.

194
00:08:44,250 --> 00:08:46,917
Because potentially, there could
be health investments or health

195
00:08:46,917 --> 00:08:49,818
behaviors that they could engage
in to improve their behavior.

196
00:08:49,818 --> 00:08:51,360
But it could also
just be that, look,

197
00:08:51,360 --> 00:08:53,140
there's actually
not much one can do.

198
00:08:53,140 --> 00:08:55,650
And so now, if we
intervene, we might actually

199
00:08:55,650 --> 00:08:56,580
make things worse.

200
00:08:56,580 --> 00:08:58,663
Because they're somebody
who actually doesn't even

201
00:08:58,663 --> 00:09:00,600
want to know what's going on.

202
00:09:00,600 --> 00:09:02,460
Often, that's actually the case.

203
00:09:02,460 --> 00:09:03,855
But instead of
just flagging that

204
00:09:03,855 --> 00:09:07,590
for motivated beliefs,
when people derive utility

205
00:09:07,590 --> 00:09:11,430
from beliefs, things
get tricky quite easily.

206
00:09:11,430 --> 00:09:13,680
And then we talked about
non-standard decision-making,

207
00:09:13,680 --> 00:09:14,700
gender discrimination.

208
00:09:14,700 --> 00:09:17,010
There's a clear case
for intervention

209
00:09:17,010 --> 00:09:20,700
if people don't get treated
the right way, as they should,

210
00:09:20,700 --> 00:09:22,830
or any other discrimination.

211
00:09:22,830 --> 00:09:25,500
And then we talked about
these other aspects here.

212
00:09:25,500 --> 00:09:27,720
Of course, then, defaults,
frames, and hudges

213
00:09:27,720 --> 00:09:29,980
are not necessarily
reasons to intervene.

214
00:09:29,980 --> 00:09:32,710
These are more like instruments
that we can use to intervene.

215
00:09:32,710 --> 00:09:35,040
You can set defaults,
or frames, or nudges

216
00:09:35,040 --> 00:09:38,520
in certain ways that
help us to make people

217
00:09:38,520 --> 00:09:43,680
make different choices in ways
that are potentially fairly

218
00:09:43,680 --> 00:09:46,380
cheap and then also in ways
that don't really necessarily

219
00:09:46,380 --> 00:09:48,360
distort behaviors of everyone.

220
00:09:48,360 --> 00:09:50,255
I'm going to get back
to that in a second.

221
00:09:50,255 --> 00:09:52,780
And when it comes to
poverty and mental health,

222
00:09:52,780 --> 00:09:54,790
these are more like
aspects that we say,

223
00:09:54,790 --> 00:09:56,730
if poverty affects
people's decision-making,

224
00:09:56,730 --> 00:10:00,420
it causes behavioral biases,
or low productivity, and so on.

225
00:10:00,420 --> 00:10:02,350
Then things like cash
transfers, [INAUDIBLE]

226
00:10:02,350 --> 00:10:05,270
might help business
[? sincerely. ?] Not sure

227
00:10:05,270 --> 00:10:06,570
that's counts as paternalism.

228
00:10:06,570 --> 00:10:08,550
That's more, like, sort
of support programs.

229
00:10:08,550 --> 00:10:11,220
And then things like
happiness and mental health,

230
00:10:11,220 --> 00:10:12,870
here you would
think about if you

231
00:10:12,870 --> 00:10:15,190
think people are
misoptimizing in certain ways,

232
00:10:15,190 --> 00:10:19,500
they sort of underuse
psychological services,

233
00:10:19,500 --> 00:10:22,050
or if they sort of
misoptimize in various ways,

234
00:10:22,050 --> 00:10:24,600
but we think they could
be experimenting more,

235
00:10:24,600 --> 00:10:27,733
they could be doing more things
to improve their own happiness.

236
00:10:27,733 --> 00:10:29,150
Perhaps there's
some ways in which

237
00:10:29,150 --> 00:10:31,530
you could push people to
experiment more and try

238
00:10:31,530 --> 00:10:32,710
to figure that out.

239
00:10:32,710 --> 00:10:35,190
Notice that I'm sort of
careful here in my wording.

240
00:10:35,190 --> 00:10:37,290
I'm saying "pushing
people to experiment

241
00:10:37,290 --> 00:10:39,690
and trying things out"
as opposed to, like,

242
00:10:39,690 --> 00:10:41,400
pushing people to
do certain things.

243
00:10:41,400 --> 00:10:44,760
For example, if we said
people should be working more

244
00:10:44,760 --> 00:10:46,478
and they work to
more or working less.

245
00:10:46,478 --> 00:10:48,270
And while it's very
hard for the government

246
00:10:48,270 --> 00:10:50,880
to actually understand
what's good and bad,

247
00:10:50,880 --> 00:10:53,220
but it could be that people
are maybe working too much.

248
00:10:53,220 --> 00:10:54,870
And you could sort of
push them towards trying

249
00:10:54,870 --> 00:10:55,620
to think about it.

250
00:10:55,620 --> 00:10:57,960
And then see that, OK, they
want to experiment and sort

251
00:10:57,960 --> 00:10:58,650
of learn more.

252
00:10:58,650 --> 00:11:01,990
Would they perhaps be
happier by doing so?

253
00:11:01,990 --> 00:11:03,180
Any questions on this?

254
00:11:09,530 --> 00:11:13,198
So then there's different
forms of paternalisms

255
00:11:13,198 --> 00:11:14,240
that you can think about.

256
00:11:14,240 --> 00:11:17,210
There's hard paternalism and
there's soft paternalism.

257
00:11:17,210 --> 00:11:20,000
And what's hard paternalism,
essentially forcing choices

258
00:11:20,000 --> 00:11:22,940
on people, changing
prices significantly,

259
00:11:22,940 --> 00:11:25,700
and mandate choices
or procedures,

260
00:11:25,700 --> 00:11:29,150
outlaw products,
including taxes.

261
00:11:29,150 --> 00:11:31,733
So you think of taxes as
sort of changing prices.

262
00:11:31,733 --> 00:11:32,900
You can make things cheaper.

263
00:11:32,900 --> 00:11:35,450
You can make things more
expensive for people.

264
00:11:35,450 --> 00:11:37,040
You can subsidize things.

265
00:11:37,040 --> 00:11:39,042
You can outlaw [INAUDIBLE].

266
00:11:41,840 --> 00:11:44,600
Hard paternalism, if you're
really significantly changing

267
00:11:44,600 --> 00:11:48,770
people's optimization
problem in pretty harsh ways

268
00:11:48,770 --> 00:11:52,640
by essentially changing
prices, potentially

269
00:11:52,640 --> 00:11:55,770
making some options
infinitely expensive.

270
00:11:55,770 --> 00:11:56,270
Right?

271
00:11:56,270 --> 00:11:58,400
It's just to say, look,
I'm just not allowing

272
00:11:58,400 --> 00:11:59,900
you to buy certain goods.

273
00:11:59,900 --> 00:12:02,990
You can think of the price as
just being infinitely expensive

274
00:12:02,990 --> 00:12:03,750
for those goods.

275
00:12:03,750 --> 00:12:05,380
So that's hard paternalism.

276
00:12:05,380 --> 00:12:08,870
Now in contrast, there's other
forms of soft paternalism.

277
00:12:08,870 --> 00:12:10,737
There's two versions of that.

278
00:12:10,737 --> 00:12:13,070
There's what's called the
libertarian paternalism, which

279
00:12:13,070 --> 00:12:16,940
is Thaler and Sunstein
are known for,

280
00:12:16,940 --> 00:12:19,550
which are essentially policies
that constructively influence

281
00:12:19,550 --> 00:12:22,460
behavior while preserving
or nearly preserving freedom

282
00:12:22,460 --> 00:12:23,100
of choice.

283
00:12:23,100 --> 00:12:24,950
But that's essentially
nudges that we

284
00:12:24,950 --> 00:12:26,520
talked about previously.

285
00:12:26,520 --> 00:12:29,820
Essentially, they're saying, you
can choose whatever you want.

286
00:12:29,820 --> 00:12:33,740
But I'm trying to influence
your behavior in some ways that

287
00:12:33,740 --> 00:12:37,400
push you towards some
options or some outcomes

288
00:12:37,400 --> 00:12:39,690
that we think are
better than others.

289
00:12:39,690 --> 00:12:40,190
Right?

290
00:12:40,190 --> 00:12:43,970
Notice that the government,
or the policymaker,

291
00:12:43,970 --> 00:12:46,310
needs to be very clear on
what's good and what's bad.

292
00:12:46,310 --> 00:12:47,630
And then, sort of try
to understand if maybe

293
00:12:47,630 --> 00:12:50,040
here's some good behavior
that we have identified,

294
00:12:50,040 --> 00:12:53,900
let's now try to push you
towards that direction.

295
00:12:53,900 --> 00:12:59,000
But since we're kind of
unsure what people want,

296
00:12:59,000 --> 00:13:01,190
we're not sort of trying
to exclude some options

297
00:13:01,190 --> 00:13:02,565
for some people
because you might

298
00:13:02,565 --> 00:13:05,460
make things worse
for some people,

299
00:13:05,460 --> 00:13:08,390
but instead just try to
push you towards that.

300
00:13:08,390 --> 00:13:09,870
Very much related
to that is what's

301
00:13:09,870 --> 00:13:11,970
called asymmetric paternalism.

302
00:13:11,970 --> 00:13:15,710
This is the term from
Camerer et al., which

303
00:13:15,710 --> 00:13:17,950
are policies that help
people who make mistakes

304
00:13:17,950 --> 00:13:22,160
while interfering minimally with
people who behave optimally.

305
00:13:22,160 --> 00:13:24,740
That is to say, you
want to be careful.

306
00:13:24,740 --> 00:13:27,350
If we see people
doing certain things,

307
00:13:27,350 --> 00:13:28,970
you want to be
careful not to make

308
00:13:28,970 --> 00:13:31,395
things worse for some people.

309
00:13:31,395 --> 00:13:33,770
Suppose there are some people,
for a share of the people,

310
00:13:33,770 --> 00:13:35,960
like, half of the population,
as behavioral agents,

311
00:13:35,960 --> 00:13:37,460
they make a bunch of mistakes.

312
00:13:37,460 --> 00:13:40,985
And half of the population
doesn't make mistakes at all.

313
00:13:40,985 --> 00:13:42,860
Now you could say and
some people are smoking

314
00:13:42,860 --> 00:13:44,902
and some people are not
smoking in each of these.

315
00:13:44,902 --> 00:13:47,168
You might say, well, the
smokers are making mistakes.

316
00:13:47,168 --> 00:13:48,960
The behavioral smokers
are making mistakes.

317
00:13:48,960 --> 00:13:49,940
So let's help them.

318
00:13:49,940 --> 00:13:51,780
So they're going to tax smoking.

319
00:13:51,780 --> 00:13:52,280
Right?

320
00:13:52,280 --> 00:13:53,947
Now what's going to
happen is, well, you

321
00:13:53,947 --> 00:13:56,300
might just improve some
outcomes for those people.

322
00:13:56,300 --> 00:13:58,590
And that might be good.

323
00:13:58,590 --> 00:14:00,987
But there will be other
people who just like smoking.

324
00:14:00,987 --> 00:14:02,820
And they're just happy,
very happy to smoke.

325
00:14:02,820 --> 00:14:04,570
And they know that
there are health risks.

326
00:14:04,570 --> 00:14:05,910
But they decide to smoke anyway.

327
00:14:05,910 --> 00:14:07,590
If you now start to
tax smoking, you're

328
00:14:07,590 --> 00:14:09,507
going to distort their
choices and make things

329
00:14:09,507 --> 00:14:10,990
worse off for them.

330
00:14:10,990 --> 00:14:14,340
So in particular,
asymmetric paternalism

331
00:14:14,340 --> 00:14:18,030
is trying to avoid that by
essentially saying, look,

332
00:14:18,030 --> 00:14:21,000
we will try to come
up with some policies

333
00:14:21,000 --> 00:14:23,340
to make things better
for some people.

334
00:14:23,340 --> 00:14:25,140
But we want to be
very careful not

335
00:14:25,140 --> 00:14:28,080
to make things worse for others,
in particular, people who

336
00:14:28,080 --> 00:14:30,740
are already optimizing, anyway.

337
00:14:30,740 --> 00:14:31,310
OK?

338
00:14:31,310 --> 00:14:33,740
And so libertarian paternalism
is a version of that.

339
00:14:33,740 --> 00:14:35,180
Or it's just very
closely related.

340
00:14:35,180 --> 00:14:39,440
Which is essentially to say, if
you preserve or merely preserve

341
00:14:39,440 --> 00:14:42,380
freedom of choice,
then it often follows

342
00:14:42,380 --> 00:14:45,620
that you're not going
to make things worse

343
00:14:45,620 --> 00:14:47,130
for people optimizing anyway.

344
00:14:47,130 --> 00:14:47,630
Right?

345
00:14:47,630 --> 00:14:49,430
Because if somebody
is optimizing anyway,

346
00:14:49,430 --> 00:14:53,060
and I'm just essentially sort
of providing these small nudges

347
00:14:53,060 --> 00:14:55,483
one way or the
other, then often,

348
00:14:55,483 --> 00:14:56,900
I'm not going to
make things worse

349
00:14:56,900 --> 00:14:58,170
for people optimizing anyway.

350
00:14:58,170 --> 00:14:59,795
That's not necessarily
always the case.

351
00:14:59,795 --> 00:15:02,540
But it's pretty much
closely related.

352
00:15:02,540 --> 00:15:04,910
I'm saying there could be
some people who just smoke--

353
00:15:04,910 --> 00:15:07,535
they have, like, five cigarettes
a day, or one cigarette a day.

354
00:15:07,535 --> 00:15:10,540
And suppose there's
two types of people.

355
00:15:10,540 --> 00:15:13,100
There's one type of
people who are essentially

356
00:15:13,100 --> 00:15:14,540
making mistakes.

357
00:15:14,540 --> 00:15:16,430
They have internalities
in the sense

358
00:15:16,430 --> 00:15:19,100
that they misunderstand
the consequences of smoking

359
00:15:19,100 --> 00:15:20,760
for their future self.

360
00:15:20,760 --> 00:15:22,600
These are people
who make mistakes.

361
00:15:22,600 --> 00:15:24,920
Now those people who
might want to help

362
00:15:24,920 --> 00:15:27,890
by increasing the price
of smoking, by taxing it,

363
00:15:27,890 --> 00:15:29,750
or in other ways.

364
00:15:29,750 --> 00:15:30,410
OK?

365
00:15:30,410 --> 00:15:32,420
But there could be other
people who essentially

366
00:15:32,420 --> 00:15:34,520
are fully aware of the
consequences of smoking

367
00:15:34,520 --> 00:15:35,930
for their future health.

368
00:15:35,930 --> 00:15:38,360
And those are optimizing.

369
00:15:38,360 --> 00:15:40,190
And they're essentially
in the optimization

370
00:15:40,190 --> 00:15:44,510
to decide that smoking
is optimal for them.

371
00:15:44,510 --> 00:15:48,020
Those people might be made
worse off by cigarette taxes.

372
00:15:48,020 --> 00:15:50,840
Because essentially, what
you do, what those taxes are,

373
00:15:50,840 --> 00:15:54,230
any other taxes, you
distort people's behaviors.

374
00:15:54,230 --> 00:15:59,300
And then you have a situation
where we have potentially

375
00:15:59,300 --> 00:16:02,000
made some people better
off, but also other people

376
00:16:02,000 --> 00:16:02,990
are worse off.

377
00:16:02,990 --> 00:16:07,550
And a lot of people would then
be upset or unhappy about this.

378
00:16:07,550 --> 00:16:10,010
Because they say, like, look,
those people who screw up

379
00:16:10,010 --> 00:16:12,437
their choices or they don't
know what they're doing,

380
00:16:12,437 --> 00:16:13,520
we're trying to help them.

381
00:16:13,520 --> 00:16:15,440
But now that comes at
the cost of others who

382
00:16:15,440 --> 00:16:16,880
actually prefer optimizing it.

383
00:16:16,880 --> 00:16:18,770
And in some sense, that
gets you in trouble.

384
00:16:18,770 --> 00:16:22,550
Because there will be unhappy
people on the other side of it,

385
00:16:22,550 --> 00:16:24,650
and potentially,
quite a few of them.

386
00:16:24,650 --> 00:16:25,850
All right.

387
00:16:25,850 --> 00:16:28,610
So what are the arguments
against paternalism?

388
00:16:28,610 --> 00:16:30,860
Or why is paternalism bad?

389
00:16:30,860 --> 00:16:34,430
You might say that
sounds pretty good.

390
00:16:34,430 --> 00:16:37,310
But, like, now, if you try
to actually implement this,

391
00:16:37,310 --> 00:16:40,570
what's bad about it?

392
00:16:40,570 --> 00:16:42,720
So one typical argument
is to say, well,

393
00:16:42,720 --> 00:16:44,310
people make mistakes.

394
00:16:44,310 --> 00:16:46,990
Well, the government
is made out of people.

395
00:16:46,990 --> 00:16:50,140
And so, how do we know
that the people who

396
00:16:50,140 --> 00:16:54,060
run the government actually
know better than people

397
00:16:54,060 --> 00:16:55,300
choose for themselves?

398
00:16:55,300 --> 00:16:59,000
And if the government
is also making mistakes,

399
00:16:59,000 --> 00:17:01,208
then, essentially
choosing for others,

400
00:17:01,208 --> 00:17:03,000
it's not clear that it
makes things better.

401
00:17:03,000 --> 00:17:05,339
Or it could make things worse.

402
00:17:05,339 --> 00:17:09,060
In particular, if you think
that the government doesn't have

403
00:17:09,060 --> 00:17:09,930
enough information.

404
00:17:09,930 --> 00:17:10,560
[INAUDIBLE]

405
00:17:10,560 --> 00:17:12,240
So I know what's best for me.

406
00:17:12,240 --> 00:17:16,020
And so therefore,
I'm best equipped

407
00:17:16,020 --> 00:17:18,180
to make decisions for myself.

408
00:17:18,180 --> 00:17:20,195
It might potentially,
in particular

409
00:17:20,195 --> 00:17:22,859
if there's heterogeneity,
it might make some people

410
00:17:22,859 --> 00:17:23,410
better off.

411
00:17:23,410 --> 00:17:25,500
But that might come at
the cost of some others.

412
00:17:25,500 --> 00:17:27,990
And then it's not clear that
on average or overall making

413
00:17:27,990 --> 00:17:28,840
things better.

414
00:17:28,840 --> 00:17:31,190
[INAUDIBLE] very
important [INAUDIBLE]

415
00:17:31,190 --> 00:17:35,490
in particular that
comes up a lot in 14.75,

416
00:17:35,490 --> 00:17:37,740
for example, which is
Political Economy, that

417
00:17:37,740 --> 00:17:38,930
will come up often.

418
00:17:38,930 --> 00:17:41,010
Because while we might
want to help people,

419
00:17:41,010 --> 00:17:42,060
want to give people
money, but then

420
00:17:42,060 --> 00:17:43,935
the government gets
essentially-- the village

421
00:17:43,935 --> 00:17:47,570
government or whatever gets
captured by elites [INAUDIBLE]

422
00:17:47,570 --> 00:17:49,540
keep the money for themselves.

423
00:17:49,540 --> 00:17:54,830
And so the more, the
more opportunity they

424
00:17:54,830 --> 00:17:58,340
get for that, [INAUDIBLE]
resources and power,

425
00:17:58,340 --> 00:18:02,750
the more opportunity they give
for misuse, and corruption,

426
00:18:02,750 --> 00:18:04,110
and so on, and so forth.

427
00:18:04,110 --> 00:18:07,630
And so that's
essentially saying that--

428
00:18:07,630 --> 00:18:10,480
so this one class up
argument is to say, well,

429
00:18:10,480 --> 00:18:13,500
even if the government wanted
to do the very best for people,

430
00:18:13,500 --> 00:18:17,528
if the government wanted
[INAUDIBLE] for its citizens,

431
00:18:17,528 --> 00:18:18,570
that would be hard to do.

432
00:18:18,570 --> 00:18:20,195
Because it doesn't
know how to do that.

433
00:18:20,195 --> 00:18:24,850
There is inherent problems with
that because of heterogeneity,

434
00:18:24,850 --> 00:18:27,650
knowledge issues, and
so on, and so forth.

435
00:18:27,650 --> 00:18:29,900
And then, there's a
second type of argument,

436
00:18:29,900 --> 00:18:34,390
which Tad just mentioned,
which is, well,

437
00:18:34,390 --> 00:18:36,880
who's there to say that
the government is actually

438
00:18:36,880 --> 00:18:39,920
benevolent and actually wants
people to be better off?

439
00:18:39,920 --> 00:18:43,490
Maybe they just want to get
rich, and so on, and so forth.

440
00:18:43,490 --> 00:18:45,610
And so now, if you
let them make choices,

441
00:18:45,610 --> 00:18:49,240
then that opens the door for
corruption and misuse of money,

442
00:18:49,240 --> 00:18:50,277
and so on, and so forth.

443
00:18:50,277 --> 00:18:52,610
And so there's sort of a
slippery slope argument to say,

444
00:18:52,610 --> 00:18:54,860
well, there's some things
in which I'm going to argue.

445
00:18:54,860 --> 00:18:56,500
For some things,
it's pretty clear

446
00:18:56,500 --> 00:19:00,535
that we can make
things better, or make

447
00:19:00,535 --> 00:19:02,910
some people, at least, better
off without hurting others.

448
00:19:02,910 --> 00:19:04,220
And there's some
cases where you say,

449
00:19:04,220 --> 00:19:06,137
let's have the government
do that or whatever,

450
00:19:06,137 --> 00:19:07,930
but set policies
that help people.

451
00:19:07,930 --> 00:19:09,305
By the way, it
doesn't have to be

452
00:19:09,305 --> 00:19:11,430
necessarily the
government [INAUDIBLE]

453
00:19:11,430 --> 00:19:12,790
company [INAUDIBLE].

454
00:19:15,398 --> 00:19:17,440
But then, [INAUDIBLE] once
we started with, like,

455
00:19:17,440 --> 00:19:18,970
some things, then they're going
to do more and more things.

456
00:19:18,970 --> 00:19:20,140
And at some point,
they'll reach a point

457
00:19:20,140 --> 00:19:21,160
where the government
actually doesn't

458
00:19:21,160 --> 00:19:22,160
know what they're doing.

459
00:19:22,160 --> 00:19:25,040
And then they make
things worse off.

460
00:19:25,040 --> 00:19:28,930
So I have a bunch of
different arguments here.

461
00:19:28,930 --> 00:19:31,473
I think we mentioned
quite a few of them.

462
00:19:31,473 --> 00:19:32,890
There's informational
issues, sort

463
00:19:32,890 --> 00:19:35,050
of not understanding
preferences, people optimizing

464
00:19:35,050 --> 00:19:37,090
anyway.

465
00:19:37,090 --> 00:19:39,610
The comments said, or
we discussed previously,

466
00:19:39,610 --> 00:19:42,800
that some people's choices
might be distorted.

467
00:19:42,800 --> 00:19:45,442
So these are issues of
heterogeneity, often.

468
00:19:45,442 --> 00:19:47,400
Then there's sort of the
government is terrible

469
00:19:47,400 --> 00:19:49,540
and wastes money, anyway.

470
00:19:49,540 --> 00:19:52,450
The government does not
like people's best interests

471
00:19:52,450 --> 00:19:53,000
in mind.

472
00:19:53,000 --> 00:19:55,030
There's regulatory capture.

473
00:19:55,030 --> 00:19:59,660
And there's also what Chad
mentioned, freedom of choice.

474
00:19:59,660 --> 00:20:01,160
Then there are some
other arguments,

475
00:20:01,160 --> 00:20:03,310
which is freedom of
choice matters, per se.

476
00:20:03,310 --> 00:20:04,920
That's a little
trickier to grasp.

477
00:20:04,920 --> 00:20:07,323
But you say, look, it's just
I want to choose for myself.

478
00:20:07,323 --> 00:20:09,490
And I don't want the
government to tell me anything.

479
00:20:09,490 --> 00:20:11,290
And that's important by itself.

480
00:20:11,290 --> 00:20:12,550
This is not about the outcome.

481
00:20:12,550 --> 00:20:15,940
This is about the freedom
of choice mattering itself.

482
00:20:15,940 --> 00:20:19,390
Then similarly, there's
arguments about people just

483
00:20:19,390 --> 00:20:20,710
don't like hard paternalism.

484
00:20:20,710 --> 00:20:22,780
They just don't like to
be told there are, like,

485
00:20:22,780 --> 00:20:27,967
certain options to be
excluded for various reasons.

486
00:20:27,967 --> 00:20:30,300
Perhaps partially because
they don't want the government

487
00:20:30,300 --> 00:20:33,510
to be powerful.

488
00:20:33,510 --> 00:20:35,490
And then, finally,
there is some sort

489
00:20:35,490 --> 00:20:38,700
of practical problems,
which is once we agree

490
00:20:38,700 --> 00:20:41,567
that we want some
paternalistic policies,

491
00:20:41,567 --> 00:20:42,900
which policies should we choose?

492
00:20:42,900 --> 00:20:45,730
And how do we know which
is better than another?

493
00:20:45,730 --> 00:20:48,052
I'm going to get back to
that in a bit as well.

494
00:20:48,052 --> 00:20:50,010
There's also practical
implementation problems,

495
00:20:50,010 --> 00:20:51,910
which is to say even if
the government wanted

496
00:20:51,910 --> 00:20:53,550
to do the right
things, the government

497
00:20:53,550 --> 00:20:56,050
is often inefficient,
and so on, and so forth.

498
00:20:56,050 --> 00:20:57,540
And so it's too costly to do.

499
00:20:57,540 --> 00:21:00,430
And practically, it's just
hard to do certain things.

500
00:21:00,430 --> 00:21:03,280
Now what are arguments in favor?

501
00:21:03,280 --> 00:21:06,890
I'm going to just
speed things up a bit.

502
00:21:06,890 --> 00:21:09,283
So the arguments in favor
are often very simple.

503
00:21:09,283 --> 00:21:11,450
Which is that people make
mistakes, people screw up,

504
00:21:11,450 --> 00:21:14,530
and we can help them
potentially screw up less.

505
00:21:14,530 --> 00:21:17,260
People often also don't
like to necessarily make

506
00:21:17,260 --> 00:21:18,430
choices for themselves.

507
00:21:18,430 --> 00:21:21,040
It could be just because
they don't like doing it.

508
00:21:21,040 --> 00:21:23,070
It could also be just
as costly to do that.

509
00:21:23,070 --> 00:21:24,820
And that's just, there
are some activities

510
00:21:24,820 --> 00:21:26,170
that are just hard to do.

511
00:21:26,170 --> 00:21:28,390
And I'd rather have somebody
regulate this for me,

512
00:21:28,390 --> 00:21:30,880
as opposed to, like,
me having to choose

513
00:21:30,880 --> 00:21:32,920
or make these choices
all for myself.

514
00:21:32,920 --> 00:21:34,780
An example would
be, suppose there's

515
00:21:34,780 --> 00:21:37,600
insurance plans or some
form of retirement plans,

516
00:21:37,600 --> 00:21:40,107
and so on, if there are some
options that are just terrible

517
00:21:40,107 --> 00:21:42,190
and they just would be
exploiting me and making me

518
00:21:42,190 --> 00:21:44,530
worse off, I could
probably figure it out

519
00:21:44,530 --> 00:21:46,510
if you gave me these
plans and I spent, like,

520
00:21:46,510 --> 00:21:48,730
a day studying them.

521
00:21:48,730 --> 00:21:50,470
But you know, I have
better stuff to do.

522
00:21:50,470 --> 00:21:53,470
And I'd rather have somebody
else screen out the bad plans

523
00:21:53,470 --> 00:21:56,650
and sort of then offer me
sort of the choices that

524
00:21:56,650 --> 00:21:58,710
try to not exploit me.

525
00:21:58,710 --> 00:22:01,632
And then, maybe, make some
choices from a limited set.

526
00:22:01,632 --> 00:22:03,090
That's not to say
I couldn't do it.

527
00:22:03,090 --> 00:22:05,173
But it's just sort of
saying it's efficient for me

528
00:22:05,173 --> 00:22:08,670
to try to deal with
these things on my own.

529
00:22:08,670 --> 00:22:10,730
Now, and then finally,
and this is, I think,

530
00:22:10,730 --> 00:22:12,230
one of the most
important arguments,

531
00:22:12,230 --> 00:22:16,410
is that there are a
lot of soft paternalism

532
00:22:16,410 --> 00:22:21,030
regimes where people can opt out
and are unaffected potentially.

533
00:22:21,030 --> 00:22:24,120
So when you go back to
the previous concerns

534
00:22:24,120 --> 00:22:26,520
that people might
have, these are often

535
00:22:26,520 --> 00:22:28,470
arguments against
hard paternalism.

536
00:22:28,470 --> 00:22:30,870
These are often things about
government wasting money,

537
00:22:30,870 --> 00:22:32,310
and so on, and so forth.

538
00:22:32,310 --> 00:22:36,900
If this is about sending
reminders or, like, phrasing

539
00:22:36,900 --> 00:22:40,860
letters in one way versus
another, in some simple way,

540
00:22:40,860 --> 00:22:43,380
or putting food or
putting apples next

541
00:22:43,380 --> 00:22:45,780
to the counter
versus candy bars,

542
00:22:45,780 --> 00:22:48,690
this is not about the government
taking over or restricting

543
00:22:48,690 --> 00:22:50,010
people's free choice.

544
00:22:50,010 --> 00:22:52,620
It's rather to say we're just
trying to make people better

545
00:22:52,620 --> 00:22:56,350
off in some simple ways to do.

546
00:22:56,350 --> 00:23:01,200
And so a lot of these
criticism or arguments

547
00:23:01,200 --> 00:23:05,810
go away once you go into
soft paternalism territory.

548
00:23:05,810 --> 00:23:08,022
And so then, at
the end of the day,

549
00:23:08,022 --> 00:23:09,730
I think it's hard to
make arguments often

550
00:23:09,730 --> 00:23:10,960
for paternalism.

551
00:23:10,960 --> 00:23:14,080
Because you need to be very
confident [INAUDIBLE],, right?

552
00:23:14,080 --> 00:23:18,130
It's much easier to argue
for soft paternalism,

553
00:23:18,130 --> 00:23:20,440
simple policies that
are not expensive,

554
00:23:20,440 --> 00:23:22,630
that arguably are not
hurting a lot of people,

555
00:23:22,630 --> 00:23:26,540
that make at least
some people better off.

556
00:23:26,540 --> 00:23:27,640
And I think-- sorry.

557
00:23:27,640 --> 00:23:29,280
This is a little
hard to see, maybe.

558
00:23:29,280 --> 00:23:31,590
But sometimes it's
actually quite hard

559
00:23:31,590 --> 00:23:34,230
to figure out whether
people are making mistakes.

560
00:23:34,230 --> 00:23:38,730
This is, for example,
a survey from Gallup

561
00:23:38,730 --> 00:23:40,480
that looks at
Americans and looks

562
00:23:40,480 --> 00:23:44,790
at people who call
themselves very or somewhat

563
00:23:44,790 --> 00:23:46,740
overweight versus
the percentage who

564
00:23:46,740 --> 00:23:50,650
are overweight and obese by some
standards by the CDC and so on.

565
00:23:50,650 --> 00:23:52,680
And so here you see
essentially that people

566
00:23:52,680 --> 00:23:56,460
tend to [INAUDIBLE] more obese.

567
00:23:56,460 --> 00:23:59,340
Because the US tends to
be quite high [INAUDIBLE]

568
00:23:59,340 --> 00:24:01,140
increasing people's
perception of that

569
00:24:01,140 --> 00:24:03,353
might be not
necessarily the case.

570
00:24:03,353 --> 00:24:04,770
Now in some ways,
like, who are we

571
00:24:04,770 --> 00:24:07,145
to sort of say that people
should change their behaviors,

572
00:24:07,145 --> 00:24:09,640
or that they're making mistakes?

573
00:24:09,640 --> 00:24:11,820
So in some situations, at
least, it's quite tricky

574
00:24:11,820 --> 00:24:14,490
to figure out what's a
mistake and what's not

575
00:24:14,490 --> 00:24:16,210
and what should we do or not.

576
00:24:16,210 --> 00:24:19,002
And, again, you get
into tricky issues

577
00:24:19,002 --> 00:24:21,210
when it comes to, in
particular, for different selves

578
00:24:21,210 --> 00:24:22,590
and self-control.

579
00:24:22,590 --> 00:24:24,867
Again, who are we
to say that somebody

580
00:24:24,867 --> 00:24:26,700
should lose weight, or
change their behavior

581
00:24:26,700 --> 00:24:30,648
in certain ways, when perhaps
that's not in their best

582
00:24:30,648 --> 00:24:32,440
interest, or that's
not what they're doing.

583
00:24:32,440 --> 00:24:34,920
Maybe they're fully optimizing.

584
00:24:34,920 --> 00:24:38,680
How would we know that
that's not the case?

585
00:24:38,680 --> 00:24:40,960
Now what are some
forms of paternalism

586
00:24:40,960 --> 00:24:43,330
that are quite popular?

587
00:24:43,330 --> 00:24:46,780
There's lots of examples of in
fact for popular paternalism.

588
00:24:46,780 --> 00:24:47,830
There's Social Security.

589
00:24:47,830 --> 00:24:49,660
There's health care
programs for retirees,

590
00:24:49,660 --> 00:24:53,650
including Medicare, which
tends to be quite popular.

591
00:24:53,650 --> 00:24:56,350
There's often restrictions
of investment menus

592
00:24:56,350 --> 00:24:58,900
and 401k savings plans,
where essentially that's

593
00:24:58,900 --> 00:25:00,520
precisely what I
was just saying,

594
00:25:00,520 --> 00:25:03,790
is where essentially just
bad plans and dominated plans

595
00:25:03,790 --> 00:25:06,070
are often just excluded.

596
00:25:06,070 --> 00:25:09,190
There's often consumer safety
regulations that the FDA has

597
00:25:09,190 --> 00:25:11,658
essentially certain
drugs cannot be offered.

598
00:25:11,658 --> 00:25:14,200
You cannot take certain drugs
the government essentially just

599
00:25:14,200 --> 00:25:17,980
sort of [INAUDIBLE] if
they're not being safe.

600
00:25:17,980 --> 00:25:19,630
There's mandatory education.

601
00:25:19,630 --> 00:25:21,400
In many places, you
have to go to school.

602
00:25:21,400 --> 00:25:27,380
You can't just sort of say
I'm just-- sort of my kid

603
00:25:27,380 --> 00:25:28,550
is learning in the forest.

604
00:25:28,550 --> 00:25:30,156
That's not an option.

605
00:25:30,156 --> 00:25:32,900
It often has to
do with protection

606
00:25:32,900 --> 00:25:35,340
of minors in some ways.

607
00:25:35,340 --> 00:25:37,730
Then there's cigarette
taxes, or sin taxes,

608
00:25:37,730 --> 00:25:38,870
as they're often called.

609
00:25:38,870 --> 00:25:40,200
But it's also hard paternalism.

610
00:25:40,200 --> 00:25:42,860
Notice that here, you're
changing people's prices.

611
00:25:42,860 --> 00:25:47,368
In many places, prostitution,
polygamy are banned.

612
00:25:47,368 --> 00:25:48,410
These are like hard bans.

613
00:25:48,410 --> 00:25:54,320
So you're just not allowed
to marry three spouses.

614
00:25:54,320 --> 00:25:56,060
Mandating face masks,
interestingly, that

615
00:25:56,060 --> 00:25:58,220
seems to be-- that's hard
paternalism, as you have,

616
00:25:58,220 --> 00:26:00,070
for example, in Cambridge.

617
00:26:00,070 --> 00:26:01,570
I think that the
punishment is going

618
00:26:01,570 --> 00:26:04,670
to be $300 if you go outside
and don't wear a face mask.

619
00:26:04,670 --> 00:26:06,890
I'm pretty sure in Cambridge
that's quite popular.

620
00:26:06,890 --> 00:26:09,840
Notice, of course, that's not
true in all states, and so on.

621
00:26:09,840 --> 00:26:11,510
So there might be
some policies that

622
00:26:11,510 --> 00:26:14,370
are very much popular
in some places,

623
00:26:14,370 --> 00:26:16,860
but not at all in others.

624
00:26:16,860 --> 00:26:20,210
There's also some
soft paternalism

625
00:26:20,210 --> 00:26:23,570
that tends to be quite popular.

626
00:26:23,570 --> 00:26:25,460
There's some places
where it's not clear

627
00:26:25,460 --> 00:26:27,360
whether it's hard
versus soft paternalism,

628
00:26:27,360 --> 00:26:29,270
or what exactly
the definition is.

629
00:26:29,270 --> 00:26:32,170
Things like banning junk food
in school vending machines,

630
00:26:32,170 --> 00:26:33,920
you can call that sort
of hard paternalism

631
00:26:33,920 --> 00:26:36,220
because you take away
options but of course, it

632
00:26:36,220 --> 00:26:40,220
doesn't really prevent you from
bringing in candy to school

633
00:26:40,220 --> 00:26:40,730
anyway.

634
00:26:40,730 --> 00:26:42,997
So why call that hard
or soft paternalism?

635
00:26:42,997 --> 00:26:43,830
I don't really care.

636
00:26:43,830 --> 00:26:47,060
But that tends to be
quite a popular policy.

637
00:26:47,060 --> 00:26:49,860
Calorie disclosure laws, which
essentially just is saying,

638
00:26:49,860 --> 00:26:53,807
OK, you have to put in some
information for our customers,

639
00:26:53,807 --> 00:26:55,140
which could be easily available.

640
00:26:55,140 --> 00:26:57,515
You could just look it
up online or the like.

641
00:26:57,515 --> 00:26:59,390
And things like default
enrollment and 401(k)

642
00:26:59,390 --> 00:27:04,560
[INAUDIBLE] all tend
to be quite popular.

643
00:27:04,560 --> 00:27:08,960
And for some goods, you
know, there's huge taxes.

644
00:27:08,960 --> 00:27:12,180
For example, cigarette taxes
tend to be extremely high.

645
00:27:12,180 --> 00:27:14,300
This is from 2019.

646
00:27:14,300 --> 00:27:18,270
Notice that there's this
huge variation across places.

647
00:27:18,270 --> 00:27:24,510
If you look at some places here,
these are some of the highest,

648
00:27:24,510 --> 00:27:27,880
such as Connecticut,
including Massachusetts,

649
00:27:27,880 --> 00:27:31,050
it's $3, $4 per
pack of cigarettes.

650
00:27:31,050 --> 00:27:34,710
In addition, there's also like
$1 from the federal government.

651
00:27:34,710 --> 00:27:41,820
Then there's some other
places, including North Dakota,

652
00:27:41,820 --> 00:27:46,920
or North Carolina, and so
on, where cigarette taxes

653
00:27:46,920 --> 00:27:49,110
are extremely low.

654
00:27:49,110 --> 00:27:50,610
So that's kind of
quite interesting,

655
00:27:50,610 --> 00:27:53,220
because it sort of
says, here's a policy

656
00:27:53,220 --> 00:27:58,140
that, depending on your view
of how people make choices,

657
00:27:58,140 --> 00:28:01,350
governments make very
different decisions.

658
00:28:01,350 --> 00:28:04,500
And presumably that has to
do with freedom of choice

659
00:28:04,500 --> 00:28:09,330
or the popularity of that choice
itself in the specific state.

660
00:28:09,330 --> 00:28:11,520
Of course, it also has
to do with, potentially,

661
00:28:11,520 --> 00:28:13,820
the party, and so on.

662
00:28:13,820 --> 00:28:16,300
Similarly, there's also
an aspiration [INAUDIBLE]

663
00:28:16,300 --> 00:28:18,850
state alcohol taxes.

664
00:28:18,850 --> 00:28:21,220
In fact, they vary hugely.

665
00:28:21,220 --> 00:28:23,890
In particular, you
find some states

666
00:28:23,890 --> 00:28:30,610
where the taxes per
gallon are $20, $30,

667
00:28:30,610 --> 00:28:34,180
and some other states where
it's essentially close to 0.

668
00:28:34,180 --> 00:28:38,110
Notice that there's also
a bunch of other blue laws

669
00:28:38,110 --> 00:28:40,900
as they're referred to, which is
things, like, on certain days,

670
00:28:40,900 --> 00:28:44,938
you're not allowed to
sell alcohol, and so on.

671
00:28:44,938 --> 00:28:46,230
Now that's kind of interesting.

672
00:28:46,230 --> 00:28:50,380
Because that's sort of
additional restrictions,

673
00:28:50,380 --> 00:28:54,510
which would tend to be quite
inefficient in various ways.

674
00:28:54,510 --> 00:28:56,310
That is to say, some
economists would

675
00:28:56,310 --> 00:28:59,370
tend to think that if there's
certain behavior that we

676
00:28:59,370 --> 00:29:03,360
don't want people to engage
in, we should tax it.

677
00:29:03,360 --> 00:29:06,030
Instead, you could
essentially sort of outlaw it

678
00:29:06,030 --> 00:29:07,170
in certain ways.

679
00:29:07,170 --> 00:29:08,670
That tends to be
quite inefficient.

680
00:29:08,670 --> 00:29:11,370
Because often, people find
ways to get it anyway.

681
00:29:11,370 --> 00:29:13,050
Plus, the government
does not get

682
00:29:13,050 --> 00:29:16,910
any taxes or revenue from that.

683
00:29:16,910 --> 00:29:20,570
Now what are some unpopular
paternalism or things

684
00:29:20,570 --> 00:29:23,870
that people don't like?

685
00:29:23,870 --> 00:29:24,970
That's exactly it.

686
00:29:24,970 --> 00:29:28,270
One of them is
helmets and seatbelts.

687
00:29:28,270 --> 00:29:31,960
So there, often there's
no externality here.

688
00:29:31,960 --> 00:29:35,320
This is essentially just
something to protect yourself.

689
00:29:35,320 --> 00:29:39,480
While they tend to be quite
unpopular in many places,

690
00:29:39,480 --> 00:29:41,950
it's not quite
clear why that is.

691
00:29:41,950 --> 00:29:44,020
So often, it's sort of
a thing that's not cool,

692
00:29:44,020 --> 00:29:46,960
or like there's some social
norms in some ways where people

693
00:29:46,960 --> 00:29:48,900
think it's not a
thing you want to do.

694
00:29:48,900 --> 00:29:52,300
So I was sort of explaining
to my 14-year-old, 15-year-old

695
00:29:52,300 --> 00:29:54,435
niece to wear a helmet
was sort of very hard.

696
00:29:54,435 --> 00:29:55,810
And then she would
say, yes, yes,

697
00:29:55,810 --> 00:29:58,960
but then drive around, ride
her bike around the corner

698
00:29:58,960 --> 00:30:01,340
and then take it off, and so on.

699
00:30:01,340 --> 00:30:06,030
So that seems to be about
something about personal trade

700
00:30:06,030 --> 00:30:08,690
offs between risk and
reward in some ways

701
00:30:08,690 --> 00:30:12,440
where the government
of some people

702
00:30:12,440 --> 00:30:16,760
might just not value the
benefits of looking cool

703
00:30:16,760 --> 00:30:20,680
or whatever sufficiently, or as
much as the individual itself.

704
00:30:20,680 --> 00:30:24,650
And then they tend to be
sort of unhappy about it.

705
00:30:24,650 --> 00:30:26,840
Of course, the alternative
way of putting this

706
00:30:26,840 --> 00:30:29,900
is that the individual itself
is underestimating the risk

707
00:30:29,900 --> 00:30:32,750
potentially or doesn't take
into account how bad it

708
00:30:32,750 --> 00:30:36,680
would be to actually have a
bicycle accident or the like.

709
00:30:36,680 --> 00:30:40,527
But that's essentially something
about people disagreeing

710
00:30:40,527 --> 00:30:42,860
with the government, or with
their parents, or the like,

711
00:30:42,860 --> 00:30:47,000
about the risk versus
reward and the cost benefits

712
00:30:47,000 --> 00:30:48,080
of certain actions.

713
00:30:48,080 --> 00:30:50,480
And there seems to
be disagreement.

714
00:30:50,480 --> 00:30:52,150
And then you could
sort of mandate that.

715
00:30:52,150 --> 00:30:53,570
But that leads to unhappiness.

716
00:30:53,570 --> 00:30:56,000
Because there's precisely
that disagreement.

717
00:30:56,000 --> 00:31:00,770
Now wearing masks is in
some sense quite similar.

718
00:31:00,770 --> 00:31:02,122
There, it's quite different.

719
00:31:02,122 --> 00:31:04,580
What's different here is that
there's externalities, right?

720
00:31:04,580 --> 00:31:08,090
So the externalities here are
that certain people benefit

721
00:31:08,090 --> 00:31:10,280
a lot from having
their shops open,

722
00:31:10,280 --> 00:31:13,250
or being able to sell things.

723
00:31:13,250 --> 00:31:15,720
And others are affected a lot.

724
00:31:15,720 --> 00:31:20,180
So in particular,
the elderly might

725
00:31:20,180 --> 00:31:24,050
be disproportionately affected
and sort of might essentially--

726
00:31:24,050 --> 00:31:28,400
the number of people who
will die, well, clearly,

727
00:31:28,400 --> 00:31:31,280
if people are not wearing
masks, or if shops,

728
00:31:31,280 --> 00:31:34,220
if people are running
outside, or meeting,

729
00:31:34,220 --> 00:31:36,330
and so on, and so forth.

730
00:31:36,330 --> 00:31:40,158
But there's others
who are really--

731
00:31:40,158 --> 00:31:41,450
for them, that's really costly.

732
00:31:41,450 --> 00:31:44,570
Because their businesses
are not open and the like.

733
00:31:44,570 --> 00:31:50,790
Now there could be
entirely rational issues

734
00:31:50,790 --> 00:31:53,790
where you might say,
well, people just

735
00:31:53,790 --> 00:31:57,022
don't care about certain
other groups of the society.

736
00:31:57,022 --> 00:31:58,980
And therefore, you want
to have your shop open.

737
00:31:58,980 --> 00:32:02,533
And you don't want to be told
that you should be doing that.

738
00:32:02,533 --> 00:32:04,950
There could be also mistakes
in the sense that people just

739
00:32:04,950 --> 00:32:07,492
don't understand, necessarily,
there are these externalities.

740
00:32:07,492 --> 00:32:09,990
They might sort of say,
well, I'm not affected.

741
00:32:09,990 --> 00:32:10,950
And I'll be just fine.

742
00:32:10,950 --> 00:32:13,860
And it might be, actually,
that the risk for themselves

743
00:32:13,860 --> 00:32:14,910
is not that high.

744
00:32:14,910 --> 00:32:17,910
But of course, there's large
externalities in others

745
00:32:17,910 --> 00:32:21,200
that they might not appreciate
or not take into account.

746
00:32:21,200 --> 00:32:21,700
Exactly.

747
00:32:21,700 --> 00:32:26,450
So Chad, Carmen is saying
speed limits, similarly,

748
00:32:26,450 --> 00:32:28,650
are similar to that.

749
00:32:28,650 --> 00:32:30,550
Because it's sort of,
like, I'm driving.

750
00:32:30,550 --> 00:32:36,330
But there's [INAUDIBLE]
in certain accidents.

751
00:32:36,330 --> 00:32:39,570
And so why is that?

752
00:32:39,570 --> 00:32:45,220
Or so how do we think
about [INAUDIBLE] say?

753
00:32:45,220 --> 00:32:48,300
So I mentioned here,
this is an example.

754
00:32:48,300 --> 00:32:54,200
So here we have [INAUDIBLE]
are quite popular.

755
00:32:54,200 --> 00:32:55,510
And why is that popular?

756
00:32:55,510 --> 00:33:00,760
And why is making certain drugs
illegal, why is that unpopular?

757
00:33:00,760 --> 00:33:02,800
What's the problem there?

758
00:33:02,800 --> 00:33:04,660
So I think there's,
when you think about,

759
00:33:04,660 --> 00:33:07,220
when you do, like,
economic analysis of that,

760
00:33:07,220 --> 00:33:10,740
is [INAUDIBLE] you can think
this might be externality.

761
00:33:10,740 --> 00:33:12,490
Externalities might
actually be much lower

762
00:33:12,490 --> 00:33:14,230
than for smoking or drinking.

763
00:33:14,230 --> 00:33:17,290
Because drunk driving
and so on is very lethal.

764
00:33:17,290 --> 00:33:19,360
There might be also
some externalities

765
00:33:19,360 --> 00:33:22,210
from smoking weed.

766
00:33:22,210 --> 00:33:26,220
But the externalities
[INAUDIBLE]

767
00:33:26,220 --> 00:33:29,170
are for drinking, well, then
there could be [INAUDIBLE]

768
00:33:29,170 --> 00:33:31,510
But, again, for internalities,
usually the solution

769
00:33:31,510 --> 00:33:34,422
would be just the typical
sin taxes, and so on.

770
00:33:34,422 --> 00:33:36,130
You can say, well, do
you want to tax it.

771
00:33:36,130 --> 00:33:36,630
Sure.

772
00:33:36,630 --> 00:33:38,505
But then you at least
would get some revenue.

773
00:33:38,505 --> 00:33:39,713
You would increase the price.

774
00:33:39,713 --> 00:33:41,740
You would get some revenue
for the government,

775
00:33:41,740 --> 00:33:44,840
plus you would lower
people's consumption.

776
00:33:44,840 --> 00:33:48,610
In addition, there's all this
stuff about locking people up

777
00:33:48,610 --> 00:33:49,510
unnecessarily.

778
00:33:49,510 --> 00:33:51,010
By making it illegal,
essentially we

779
00:33:51,010 --> 00:33:53,630
create a large
prisoner population,

780
00:33:53,630 --> 00:33:57,770
and so on, which surely
is [INAUDIBLE] society.

781
00:33:57,770 --> 00:34:00,610
So once you sort of have
this policy [INAUDIBLE],,

782
00:34:00,610 --> 00:34:02,260
there's actually
not a lot of reasons

783
00:34:02,260 --> 00:34:05,890
for outlawing how
these sort of tend

784
00:34:05,890 --> 00:34:10,732
to be gut reactions or
some sort of judgments

785
00:34:10,732 --> 00:34:11,940
that were made at some point.

786
00:34:11,940 --> 00:34:14,290
And there's some lobby
arguing in favor of alcohol

787
00:34:14,290 --> 00:34:17,530
and maybe less of a lobby
arguing for other drugs.

788
00:34:17,530 --> 00:34:20,469
But as we say, the sentiment
has very much changed.

789
00:34:20,469 --> 00:34:22,719
And it seems to me, sooner
or later, pretty much

790
00:34:22,719 --> 00:34:27,530
all states are moving
towards legalization

791
00:34:27,530 --> 00:34:31,969
and that leads to, essentially,
[INAUDIBLE] government

792
00:34:31,969 --> 00:34:35,335
revenues as well.

793
00:34:35,335 --> 00:34:37,460
So I think one argument
perhaps that one could make

794
00:34:37,460 --> 00:34:40,370
is sort of these
gateway-drug arguments, which

795
00:34:40,370 --> 00:34:43,260
I don't think there's actually
that much evidence for that.

796
00:34:43,260 --> 00:34:46,940
But you say, well, actually,
marijuana might not

797
00:34:46,940 --> 00:34:48,679
be so bad by itself,
but perhaps it

798
00:34:48,679 --> 00:34:51,080
leads to other
[? worries. ?] It's not

799
00:34:51,080 --> 00:34:54,560
obvious at all that
legalization of that-- so A,

800
00:34:54,560 --> 00:34:57,139
it's not obvious that that's
necessarily true, but B,

801
00:34:57,139 --> 00:35:00,530
it's not obvious that making
it legal versus illegal changes

802
00:35:00,530 --> 00:35:01,520
that rationale.

803
00:35:01,520 --> 00:35:04,370
Because when it's
illegal, lots of people

804
00:35:04,370 --> 00:35:05,510
are using drugs anyway.

805
00:35:05,510 --> 00:35:07,302
And maybe it makes it
even more [INAUDIBLE]

806
00:35:07,302 --> 00:35:10,290
and maybe more
dangerous, potentially.

807
00:35:10,290 --> 00:35:12,020
But I think it's
important to realize

808
00:35:12,020 --> 00:35:14,990
that some of these policies
are very much, in some ways,

809
00:35:14,990 --> 00:35:19,250
arbitrary, based on some
legacy or some lobbies

810
00:35:19,250 --> 00:35:20,670
and so on and so forth.

811
00:35:20,670 --> 00:35:22,695
And what's important
here is then

812
00:35:22,695 --> 00:35:26,540
that the popularity
itself then also will

813
00:35:26,540 --> 00:35:27,973
affect policies in some way.

814
00:35:27,973 --> 00:35:29,390
So like, [INAUDIBLE],,
as you say,

815
00:35:29,390 --> 00:35:31,820
the public sentiment has
shifted in various [INAUDIBLE]

816
00:35:31,820 --> 00:35:35,960
some ways stuff will become
legal as a result of that.

817
00:35:35,960 --> 00:35:38,270
And once you do it in some
states and things seem fine

818
00:35:38,270 --> 00:35:40,100
and the states make a
lot of money from it,

819
00:35:40,100 --> 00:35:44,940
a lot of other states eventually
will follow suit, in part

820
00:35:44,940 --> 00:35:49,190
because it's also much
harder to enforce laws

821
00:35:49,190 --> 00:35:50,900
if you have a bunch
of surrounding states

822
00:35:50,900 --> 00:35:52,985
where you can buy stuff legally.

823
00:35:52,985 --> 00:35:54,860
And that makes it, of
course, trickier, then,

824
00:35:54,860 --> 00:35:58,280
to enforce the law anyway.

825
00:35:58,280 --> 00:36:01,700
I have some other unpopular
examples [INAUDIBLE]

826
00:36:01,700 --> 00:36:08,120
unpopular-- paternalism, junk
food bans, junk food taxes.

827
00:36:08,120 --> 00:36:10,530
And some of these things,
again, are changing over time.

828
00:36:10,530 --> 00:36:15,170
So it may well be that
things are different

829
00:36:15,170 --> 00:36:17,000
in 10, 20 years from now.

830
00:36:17,000 --> 00:36:21,800
Gambling laws, pornography
laws, other kinds of sin taxes.

831
00:36:21,800 --> 00:36:23,850
Mandating face masks--
and again, in some states,

832
00:36:23,850 --> 00:36:25,350
they're just really,
really popular.

833
00:36:25,350 --> 00:36:27,770
And people are on the
streets protesting

834
00:36:27,770 --> 00:36:33,290
against it, which is quite
interesting [INAUDIBLE]

835
00:36:33,290 --> 00:36:34,190
might be.

836
00:36:34,190 --> 00:36:38,480
But often it is the case that
there are some forms of--

837
00:36:38,480 --> 00:36:42,380
that there's real economic
interest behind that, often.

838
00:36:42,380 --> 00:36:44,180
And perhaps,
[? so ?] again, that

839
00:36:44,180 --> 00:36:46,190
could be very much
fully rational

840
00:36:46,190 --> 00:36:48,260
and could also be that
people might misunderstand

841
00:36:48,260 --> 00:36:52,010
the externalities
that are at play.

842
00:36:52,010 --> 00:36:54,890
There's some other kinds
of hard paternalism that's

843
00:36:54,890 --> 00:36:56,990
also kind of unpopular and
also not a particularly

844
00:36:56,990 --> 00:36:59,000
efficient thing to
do, which is things

845
00:36:59,000 --> 00:37:01,593
like the ban on
16-ounce sodas, which

846
00:37:01,593 --> 00:37:04,010
is kind of like a weird thing
to do if you think about it,

847
00:37:04,010 --> 00:37:08,690
because you can just
buy smaller bottles

848
00:37:08,690 --> 00:37:14,470
or the like, which tends to
be quite inefficient overall.

849
00:37:14,470 --> 00:37:16,940
Here's a sign for seatbelts.

850
00:37:19,970 --> 00:37:22,910
Now, OK, so here's
the legalization,

851
00:37:22,910 --> 00:37:26,120
which also there is massive
variation across states.

852
00:37:26,120 --> 00:37:28,160
Now, there's some nudges.

853
00:37:28,160 --> 00:37:30,380
And some policies
have massive impacts.

854
00:37:30,380 --> 00:37:31,950
And we studied at
least some of this.

855
00:37:31,950 --> 00:37:35,380
For example, this was
automatic enrollment.

856
00:37:35,380 --> 00:37:38,780
There is essentially a
huge swing towards that.

857
00:37:38,780 --> 00:37:43,100
Essentially lots of companies
and so on are now using that.

858
00:37:43,100 --> 00:37:48,200
And the fraction of people who
are participating in retirement

859
00:37:48,200 --> 00:37:51,140
plans is way higher than
the fraction of people when

860
00:37:51,140 --> 00:37:52,700
there's voluntary [INAUDIBLE].

861
00:37:52,700 --> 00:37:57,380
There's other policies which
are quite simple policies which

862
00:37:57,380 --> 00:38:03,625
are things trying to get
people to appear in court.

863
00:38:03,625 --> 00:38:10,830
So essentially
lots of people who

864
00:38:10,830 --> 00:38:14,820
have to appear in court for,
like, initial court hearings

865
00:38:14,820 --> 00:38:16,650
tend to not show up
in those meetings.

866
00:38:16,650 --> 00:38:18,720
And as you can imagine,
that's not a great idea.

867
00:38:18,720 --> 00:38:25,020
Because that essentially leads
to really bad consequences.

868
00:38:25,020 --> 00:38:26,760
Now, what you can
essentially just do

869
00:38:26,760 --> 00:38:29,763
is try to do very
simple policies.

870
00:38:29,763 --> 00:38:32,055
And there's a very nice paper
by [INAUDIBLE] and others

871
00:38:32,055 --> 00:38:33,120
that looks at this.

872
00:38:33,120 --> 00:38:35,370
You can sort of
change the forms.

873
00:38:35,370 --> 00:38:37,350
You can phrase
things differently.

874
00:38:37,350 --> 00:38:41,200
That essentially tries to nudge
people in the right direction.

875
00:38:41,200 --> 00:38:43,735
You can also do things like
text message reminders.

876
00:38:43,735 --> 00:38:46,110
And the text message literally
just says, your court date

877
00:38:46,110 --> 00:38:49,200
is on May 22 at 9:00 AM.

878
00:38:49,200 --> 00:38:51,000
Please show up.

879
00:38:51,000 --> 00:38:52,830
And you remind
people to do that,

880
00:38:52,830 --> 00:38:55,260
and what you get
essentially is that you

881
00:38:55,260 --> 00:38:58,560
get large decreases in
timely court appearances--

882
00:38:58,560 --> 00:39:02,220
or non-appearances, or
the fraction of people

883
00:39:02,220 --> 00:39:08,040
who don't show up in court,
to their own court hearing,

884
00:39:08,040 --> 00:39:12,530
decreases a lot by using
very, very simple policies.

885
00:39:12,530 --> 00:39:15,360
Once you do those kinds
of policies, presumably,

886
00:39:15,360 --> 00:39:18,030
[INAUDIBLE] not much to
say against this policy

887
00:39:18,030 --> 00:39:22,215
in the sense that you
might say, well, so A,

888
00:39:22,215 --> 00:39:24,600
you're not restricting
anybody's choices.

889
00:39:24,600 --> 00:39:26,580
You're also not really
making anybody worse

890
00:39:26,580 --> 00:39:28,980
off by sending some
text message reminders.

891
00:39:28,980 --> 00:39:31,015
There might be some
small inconveniences

892
00:39:31,015 --> 00:39:32,640
of receiving these
text messages, which

893
00:39:32,640 --> 00:39:35,130
are kind of annoying and
spam for some people.

894
00:39:35,130 --> 00:39:38,340
But overall, nobody's
really made off by this.

895
00:39:38,340 --> 00:39:40,240
It's a very cheap thing to do.

896
00:39:40,240 --> 00:39:41,700
So there's not a
lot of arguments

897
00:39:41,700 --> 00:39:43,735
against such policies,
where you say, look,

898
00:39:43,735 --> 00:39:45,610
here's a policy that's
extremely cheap to do.

899
00:39:45,610 --> 00:39:46,860
Some people are benefiting.

900
00:39:46,860 --> 00:39:48,990
Presumably showing up in
court is a good thing--

901
00:39:48,990 --> 00:39:50,670
for your court hearing.

902
00:39:50,670 --> 00:39:53,430
And it's cheap thing to do.

903
00:39:53,430 --> 00:39:56,700
And the benefits are
potentially large.

904
00:39:56,700 --> 00:39:59,070
Because now we
prevent some people

905
00:39:59,070 --> 00:40:01,140
from essentially
ending up in jail

906
00:40:01,140 --> 00:40:04,770
for too long or for
unnecessary time

907
00:40:04,770 --> 00:40:06,870
because they failed to
appear to their hearings.

908
00:40:09,782 --> 00:40:11,490
But so then I think
it's also important--

909
00:40:11,490 --> 00:40:14,032
so there's some examples that
essentially show large effects,

910
00:40:14,032 --> 00:40:16,320
including automatic
enrollment, including things

911
00:40:16,320 --> 00:40:19,740
like text messages
of some interventions

912
00:40:19,740 --> 00:40:22,350
to induce court appearances.

913
00:40:22,350 --> 00:40:25,230
But it's important to understand
that not everything always

914
00:40:25,230 --> 00:40:25,770
works.

915
00:40:25,770 --> 00:40:27,750
What you see often,
or what I've shown

916
00:40:27,750 --> 00:40:32,530
you are examples of successful
examples of stuff that works.

917
00:40:32,530 --> 00:40:35,000
But there are some examples
that just don't work at all.

918
00:40:35,000 --> 00:40:37,510
So here's an example of a
clinical trial, which is called

919
00:40:37,510 --> 00:40:41,020
the HeartStrong
randomized clinical trial,

920
00:40:41,020 --> 00:40:42,978
which gave people
essentially a kitchen

921
00:40:42,978 --> 00:40:45,270
sink of things that we think
are all good things to do.

922
00:40:45,270 --> 00:40:47,070
This is stuff like
[? reminders, ?]

923
00:40:47,070 --> 00:40:49,380
financial incentives,
social support,

924
00:40:49,380 --> 00:40:52,560
and so on and so forth,
for people, essentially,

925
00:40:52,560 --> 00:40:55,230
who had heart attacks,
by giving them

926
00:40:55,230 --> 00:40:58,920
wireless bottles, lottery-based
incentives, and again,

927
00:40:58,920 --> 00:41:04,020
social support, having friends
and family to help them.

928
00:41:04,020 --> 00:41:07,820
And essentially that did
exactly nothing for people

929
00:41:07,820 --> 00:41:09,850
and had no effect whatsoever.

930
00:41:09,850 --> 00:41:12,810
So it's not to say that we
think everything works, always,

931
00:41:12,810 --> 00:41:15,480
and we should just always nudge
people and provide incentives

932
00:41:15,480 --> 00:41:16,810
and so on and so forth.

933
00:41:16,810 --> 00:41:20,265
And something that might
work but in fact there's

934
00:41:20,265 --> 00:41:22,140
no reason to think that
it might always work.

935
00:41:22,140 --> 00:41:26,830
In some situations, this might
be helpful and others not.

936
00:41:26,830 --> 00:41:30,160
More concerningly there's
also a very recent paper

937
00:41:30,160 --> 00:41:34,060
from Stefano DellaVigna that
just came out like a week ago

938
00:41:34,060 --> 00:41:37,605
that looks at academic
studies and nudge units.

939
00:41:37,605 --> 00:41:38,980
That's to say what
they're trying

940
00:41:38,980 --> 00:41:41,650
to do is look at treatment
effects in academic studies

941
00:41:41,650 --> 00:41:44,050
that are published in
papers which are papers

942
00:41:44,050 --> 00:41:48,400
like the stuff that we
saw on default effects

943
00:41:48,400 --> 00:41:52,135
and so on, which is Carroll et
al is the 401(k) savings one,

944
00:41:52,135 --> 00:41:56,200
here is the study on FAFSA
involvement that I showed you,

945
00:41:56,200 --> 00:42:00,210
there's some
changing many orders

946
00:42:00,210 --> 00:42:02,440
in buffet line for
healthier food consumption.

947
00:42:02,440 --> 00:42:04,380
So there's a bunch of
papers that usually you

948
00:42:04,380 --> 00:42:06,630
would teach in a typical class.

949
00:42:06,630 --> 00:42:09,900
And what you see here
is, on the y-axis,

950
00:42:09,900 --> 00:42:12,300
you could see treatment
effects and percentage points.

951
00:42:12,300 --> 00:42:15,150
On the x-axis here you see
the control group tick up,

952
00:42:15,150 --> 00:42:17,100
which is to say, if
you don't do anything,

953
00:42:17,100 --> 00:42:18,990
what does the control group do?

954
00:42:18,990 --> 00:42:20,787
And what do you see is
pretty large effects

955
00:42:20,787 --> 00:42:22,620
in many of these studies
that are published.

956
00:42:22,620 --> 00:42:24,910
There are some questions on this
is like U-shaped or inverse U

957
00:42:24,910 --> 00:42:25,680
or whatever.

958
00:42:25,680 --> 00:42:27,373
Let's ignore that for a second.

959
00:42:27,373 --> 00:42:29,790
But overall what you see is
pretty large treatment effects

960
00:42:29,790 --> 00:42:30,810
in these situations.

961
00:42:33,430 --> 00:42:38,260
And this is like 26 trials
and 71 different nudges here.

962
00:42:38,260 --> 00:42:41,200
When you instead look
at nudge units-- so

963
00:42:41,200 --> 00:42:42,100
what are nudge units?

964
00:42:42,100 --> 00:42:43,600
These are essentially,
for example,

965
00:42:43,600 --> 00:42:49,010
the UK government has what's
called the BIT, the Behavioral

966
00:42:49,010 --> 00:42:52,570
Insights Team, which essentially
is a team inside the UK

967
00:42:52,570 --> 00:42:55,900
government that tries to
support or nudge people

968
00:42:55,900 --> 00:42:57,030
to make better choices.

969
00:42:57,030 --> 00:43:00,340
And that's very much coming
from the nudge movement,

970
00:43:00,340 --> 00:43:03,160
from Thaler and so on,
that sort of advised the UK

971
00:43:03,160 --> 00:43:05,080
government to set this up.

972
00:43:05,080 --> 00:43:06,708
And they're now nudging away.

973
00:43:06,708 --> 00:43:09,250
Because the government is really
happy about cheap stuff that

974
00:43:09,250 --> 00:43:12,960
doesn't cost very much but
potentially has large benefits.

975
00:43:12,960 --> 00:43:17,810
Now what they then
also do, often,

976
00:43:17,810 --> 00:43:20,840
is they evaluate their
interventions in part

977
00:43:20,840 --> 00:43:23,390
because they try to figure
out what works best.

978
00:43:23,390 --> 00:43:25,190
Now, when you look at
these interventions,

979
00:43:25,190 --> 00:43:28,730
you see essentially that
the treatment effects

980
00:43:28,730 --> 00:43:30,650
in those kinds of
interventions tend

981
00:43:30,650 --> 00:43:32,108
to be much larger
than on the left.

982
00:43:32,108 --> 00:43:33,608
If you at the right,
there's a bunch

983
00:43:33,608 --> 00:43:36,050
of stuff, a massive bunch of
points that are pretty high,

984
00:43:36,050 --> 00:43:37,378
above 10%, 20%.

985
00:43:37,378 --> 00:43:39,170
When you look to the
right, essentially you

986
00:43:39,170 --> 00:43:42,470
see almost everything
is between 0% and 10%.

987
00:43:42,470 --> 00:43:45,380
A lot of it seems to be not
statistically significantly

988
00:43:45,380 --> 00:43:48,110
different from 0.

989
00:43:48,110 --> 00:43:51,050
So that should give us some
pause in the sense of we

990
00:43:51,050 --> 00:43:54,620
should be careful in necessarily
recommending a bunch of stuff

991
00:43:54,620 --> 00:43:56,750
to government and saying
we should always nudge

992
00:43:56,750 --> 00:43:59,300
and so on and so forth.

993
00:43:59,300 --> 00:44:00,330
What's going on here?

994
00:44:00,330 --> 00:44:03,318
Well, some of this could
be about publication bias.

995
00:44:03,318 --> 00:44:05,735
Essentially there's a bunch
of studies that are being run,

996
00:44:05,735 --> 00:44:07,520
and what's being
published is essentially

997
00:44:07,520 --> 00:44:11,052
stuff that's highly significant
and flashy and important.

998
00:44:11,052 --> 00:44:13,010
There could be also issues
with implementation,

999
00:44:13,010 --> 00:44:14,960
that maybe what's
done with researchers

1000
00:44:14,960 --> 00:44:17,970
is sort of implemented
better, potentially,

1001
00:44:17,970 --> 00:44:20,010
and maybe what's
done at scale is

1002
00:44:20,010 --> 00:44:23,050
just harder to do in nudge units
and harder to do carefully.

1003
00:44:23,050 --> 00:44:25,068
And then the effects
might be lower.

1004
00:44:25,068 --> 00:44:26,610
Not to say that
governments shouldn't

1005
00:44:26,610 --> 00:44:30,120
nudge in a sense of,
if there are effects

1006
00:44:30,120 --> 00:44:33,240
of the order of magnitude
something like 5%,

1007
00:44:33,240 --> 00:44:35,970
percentage points, and
that's really cheap to do,

1008
00:44:35,970 --> 00:44:37,710
you should probably
do it anyway.

1009
00:44:37,710 --> 00:44:40,002
But it's just the same, nudges
are not the magic bullet

1010
00:44:40,002 --> 00:44:42,043
that sort of change
everything and the government

1011
00:44:42,043 --> 00:44:43,313
should always engage on them.

1012
00:44:43,313 --> 00:44:44,730
And it's quite
important, in fact,

1013
00:44:44,730 --> 00:44:46,230
for the governments,
or governments,

1014
00:44:46,230 --> 00:44:48,910
or any sort of
these policy units.

1015
00:44:48,910 --> 00:44:51,323
And firms, by the way,
do the exact same thing

1016
00:44:51,323 --> 00:44:53,740
to try and figure out what
works and what doesn't, and not

1017
00:44:53,740 --> 00:44:57,130
just rely on the academic
evidence that might sort of be

1018
00:44:57,130 --> 00:45:00,010
more proof-of-concept rather
than sort of something that

1019
00:45:00,010 --> 00:45:02,200
evaluates cost-effectiveness.

1020
00:45:02,200 --> 00:45:03,460
Any questions on this?

1021
00:45:12,420 --> 00:45:15,090
Now, another question--
this is a little blurry--

1022
00:45:15,090 --> 00:45:18,907
another question you
might ask, and you

1023
00:45:18,907 --> 00:45:20,490
might have seen these
kinds of letters

1024
00:45:20,490 --> 00:45:23,220
before, whether everybody
wants to be nudged.

1025
00:45:23,220 --> 00:45:27,600
And so what there's a very
famous study by Allcott

1026
00:45:27,600 --> 00:45:31,740
and others that looked
at energy comparisons.

1027
00:45:31,740 --> 00:45:35,090
It's a study done with Opower,
the company, that essentially

1028
00:45:35,090 --> 00:45:37,340
tries to reduce energy usage.

1029
00:45:37,340 --> 00:45:39,050
And the way this
works is what you do

1030
00:45:39,050 --> 00:45:42,840
is you send people
these kinds of letters.

1031
00:45:42,840 --> 00:45:48,180
And I'm getting these letters
too that sort of say, here's

1032
00:45:48,180 --> 00:45:51,930
your energy consumption in
the last month, the last year.

1033
00:45:51,930 --> 00:45:54,383
Here is all of your neighbors.

1034
00:45:54,383 --> 00:45:55,425
So it's your comparisons.

1035
00:45:55,425 --> 00:45:58,480
You see the blue bar is
like how much did you use

1036
00:45:58,480 --> 00:46:01,230
and the gray bar
is like how much

1037
00:46:01,230 --> 00:46:02,772
did all of your neighbors use.

1038
00:46:02,772 --> 00:46:04,230
And then there's
the green bar that

1039
00:46:04,230 --> 00:46:07,390
sort of says how much do
efficient neighbors use.

1040
00:46:07,390 --> 00:46:10,960
And then there's smileys, where
you have two smileys if you're

1041
00:46:10,960 --> 00:46:12,850
doing great, one smiley
if you're doing good,

1042
00:46:12,850 --> 00:46:14,800
like this fellow here.

1043
00:46:14,800 --> 00:46:16,330
And then there's
no smiley, I guess,

1044
00:46:16,330 --> 00:46:19,180
if you're not doing well.

1045
00:46:19,180 --> 00:46:21,010
Sometimes there's
also like a sad smiley

1046
00:46:21,010 --> 00:46:23,790
if you're using lots
of energy and so on.

1047
00:46:23,790 --> 00:46:26,520
And so then essentially
what this is trying to do

1048
00:46:26,520 --> 00:46:29,100
is trying to provide some
social comparison here.

1049
00:46:29,100 --> 00:46:30,840
It's a very cheap policy to do.

1050
00:46:30,840 --> 00:46:35,670
In particular, when you're
sending energy bills anyway,

1051
00:46:35,670 --> 00:46:38,440
it's very cheap to add this
at the top, where essentially

1052
00:46:38,440 --> 00:46:40,440
you're sending people
energy bills anyway.

1053
00:46:40,440 --> 00:46:42,300
And now you're adding
this social comparison

1054
00:46:42,300 --> 00:46:45,763
to try to nudge people
to use less energy.

1055
00:46:45,763 --> 00:46:47,930
There's a question of why
the company might do that.

1056
00:46:47,930 --> 00:46:50,670
But that's a separate issue.

1057
00:46:50,670 --> 00:46:56,790
And what you see here is, when
you look at treatment effects,

1058
00:46:56,790 --> 00:46:58,980
there tends to be clearer
treatment effects--

1059
00:46:58,980 --> 00:47:02,370
at least some treatment
effects of providing

1060
00:47:02,370 --> 00:47:07,710
those kinds of information
reduces overall energy

1061
00:47:07,710 --> 00:47:11,450
consumption in RCTs
that people have.

1062
00:47:11,450 --> 00:47:14,750
Now, one question you might
ask is, well, shouldn't we

1063
00:47:14,750 --> 00:47:15,620
then always do this?

1064
00:47:15,620 --> 00:47:17,520
So what's the reason
not to do that?

1065
00:47:17,520 --> 00:47:22,780
And it turns out [INAUDIBLE] not
everybody wants to be nudged.

1066
00:47:22,780 --> 00:47:25,300
So here's an intervention
where now, again,

1067
00:47:25,300 --> 00:47:28,030
Allcott and co-author
have now, in fact,

1068
00:47:28,030 --> 00:47:32,800
asked people about, would you
like to receive these letters

1069
00:47:32,800 --> 00:47:34,540
or would you like to
receive these kinds

1070
00:47:34,540 --> 00:47:35,950
of social comparisons?

1071
00:47:35,950 --> 00:47:40,210
And what they ask is what's
your positive willingness

1072
00:47:40,210 --> 00:47:41,140
to pay for it.

1073
00:47:41,140 --> 00:47:44,890
So would you like to have a
dollar, versus these reports?

1074
00:47:44,890 --> 00:47:47,380
So a dollar and
no reports versus

1075
00:47:47,380 --> 00:47:48,860
just receiving the reports.

1076
00:47:48,860 --> 00:47:50,920
What is people's
willingness to pay?

1077
00:47:50,920 --> 00:47:52,450
But they also ask
their willingness

1078
00:47:52,450 --> 00:47:54,530
to pay not to receive
these reports.

1079
00:47:54,530 --> 00:47:57,450
That is to say, I
could either tell you--

1080
00:47:57,450 --> 00:48:06,460
I would ask you, you receive
zero dollars and no reports

1081
00:48:06,460 --> 00:48:08,320
or you receive a
dollar and the reports.

1082
00:48:08,320 --> 00:48:11,050
And some people turn that
down, which essentially

1083
00:48:11,050 --> 00:48:14,650
implies that they'd rather have
a dollar less and no reports.

1084
00:48:14,650 --> 00:48:17,980
They're willing to pay
not to receive reports.

1085
00:48:17,980 --> 00:48:19,762
And this is what you
see here on the left,

1086
00:48:19,762 --> 00:48:21,220
is essentially
people's willingness

1087
00:48:21,220 --> 00:48:22,520
to pay for home reports.

1088
00:48:22,520 --> 00:48:24,020
So there are quite
a few people here

1089
00:48:24,020 --> 00:48:25,940
on the right who have
positive willingness to pay.

1090
00:48:25,940 --> 00:48:28,070
They think it's kind of useful,
it's kind of interesting.

1091
00:48:28,070 --> 00:48:29,987
Maybe it helps them
reduce energy consumption.

1092
00:48:29,987 --> 00:48:32,230
Maybe that motivates
them in certain ways.

1093
00:48:32,230 --> 00:48:34,390
But some people are just
saying, essentially,

1094
00:48:34,390 --> 00:48:36,587
I do not want these reports.

1095
00:48:36,587 --> 00:48:38,420
And the reason being
that, essentially, it's

1096
00:48:38,420 --> 00:48:41,870
kind of just annoying to
get nagged all the time.

1097
00:48:41,870 --> 00:48:45,270
Maybe in some cases you just
don't have a lot of choice.

1098
00:48:45,270 --> 00:48:48,140
And in some cases maybe you
just like to use lots of energy

1099
00:48:48,140 --> 00:48:49,230
for whatever reason.

1100
00:48:49,230 --> 00:48:51,230
And it's really annoying
to be told all the time

1101
00:48:51,230 --> 00:48:53,525
that your neighbor is
really better than you are.

1102
00:48:53,525 --> 00:48:56,210
It makes people
uncomfortable and makes them

1103
00:48:56,210 --> 00:48:58,327
unhappy and so on and so forth.

1104
00:48:58,327 --> 00:48:59,660
Maybe it makes them feel guilty.

1105
00:48:59,660 --> 00:49:01,520
Who knows?

1106
00:49:01,520 --> 00:49:03,950
Maybe they just don't
want to think about it.

1107
00:49:03,950 --> 00:49:06,320
But importantly, there
are some people who

1108
00:49:06,320 --> 00:49:08,570
just do not want to be nudged.

1109
00:49:08,570 --> 00:49:10,520
And I think that's important.

1110
00:49:10,520 --> 00:49:13,400
Because when you think
back about the libertarian

1111
00:49:13,400 --> 00:49:16,520
paternalism or the soft
paternalism criteria,

1112
00:49:16,520 --> 00:49:19,910
asymmetric paternalism
that we talked about, well,

1113
00:49:19,910 --> 00:49:21,740
what the asymmetric
paternalism would say

1114
00:49:21,740 --> 00:49:24,710
is we should not make
anybody worse off.

1115
00:49:24,710 --> 00:49:26,240
So now if we send
all these letters

1116
00:49:26,240 --> 00:49:29,360
and make, actually,
people uncomfortable,

1117
00:49:29,360 --> 00:49:32,840
then in fact we are potentially
making people worse off.

1118
00:49:32,840 --> 00:49:36,790
And you can think about some
other types of interventions.

1119
00:49:36,790 --> 00:49:39,620
If you [INAUDIBLE] letters
from MIT all the time,

1120
00:49:39,620 --> 00:49:42,644
how are you doing
compared to classmates,

1121
00:49:42,644 --> 00:49:47,960
in terms of study efforts,
in terms of problem sets,

1122
00:49:47,960 --> 00:49:52,970
[INAUDIBLE] and so on,
potentially that would likely

1123
00:49:52,970 --> 00:49:56,120
increase study efforts in some
ways and would perhaps increase

1124
00:49:56,120 --> 00:49:57,530
learning in some ways.

1125
00:49:57,530 --> 00:49:59,990
But it's not clear at all that
we would increase welfare.

1126
00:49:59,990 --> 00:50:01,740
Because maybe everybody
would be even more

1127
00:50:01,740 --> 00:50:03,590
stressed and everybody
would sort of

1128
00:50:03,590 --> 00:50:07,730
be less happy and so on and so
forth, at least in some cases.

1129
00:50:12,220 --> 00:50:15,310
Yeah, actually so Jenna
is asking a question

1130
00:50:15,310 --> 00:50:17,140
about who's unhappy about this.

1131
00:50:24,020 --> 00:50:27,490
So the asymmetric
paternalism is just

1132
00:50:27,490 --> 00:50:34,630
to say suppose there are
some people who are just very

1133
00:50:34,630 --> 00:50:37,270
happy with their choice
about using a lot of energy,

1134
00:50:37,270 --> 00:50:39,800
and they just don't
want to be nudged.

1135
00:50:39,800 --> 00:50:41,850
They're fully rationally
choosing this.

1136
00:50:41,850 --> 00:50:43,870
Maybe they know, deep
down in their heart,

1137
00:50:43,870 --> 00:50:47,440
that it's bad to ruin the planet
and climate and so on and so

1138
00:50:47,440 --> 00:50:48,140
forth.

1139
00:50:48,140 --> 00:50:49,765
But they really made
a choice that they

1140
00:50:49,765 --> 00:50:51,130
want to use lots of energy.

1141
00:50:51,130 --> 00:50:52,750
Those people are
not the people who

1142
00:50:52,750 --> 00:50:54,550
want to receive any letters.

1143
00:50:54,550 --> 00:50:58,582
So asymmetric paternalism
would say we should just

1144
00:50:58,582 --> 00:50:59,290
leave them alone.

1145
00:50:59,290 --> 00:51:01,930
They should just be doing
whatever they want to do.

1146
00:51:01,930 --> 00:51:04,510
And us sending them letters
to make them feel bad

1147
00:51:04,510 --> 00:51:06,520
if they don't even change
their behavior anyway,

1148
00:51:06,520 --> 00:51:08,440
or even if they would
change their behavior,

1149
00:51:08,440 --> 00:51:09,502
that's not for us to say.

1150
00:51:09,502 --> 00:51:10,960
And we should just
leave them alone

1151
00:51:10,960 --> 00:51:13,422
because we make them worse
off by receiving these letters

1152
00:51:13,422 --> 00:51:15,880
by their revealed preference
to say they are willing to pay

1153
00:51:15,880 --> 00:51:18,420
not to receive these letters.

1154
00:51:18,420 --> 00:51:21,510
Now, there could be some
situations where people

1155
00:51:21,510 --> 00:51:25,060
who have potentially
large treatment effects,

1156
00:51:25,060 --> 00:51:26,790
these are people who
use lots of energy

1157
00:51:26,790 --> 00:51:28,470
and they just don't
know about it.

1158
00:51:28,470 --> 00:51:30,630
And those, you want
to sort of send them

1159
00:51:30,630 --> 00:51:32,070
letters and push them.

1160
00:51:32,070 --> 00:51:34,940
But I think the whole point
of the asymmetric paternalism

1161
00:51:34,940 --> 00:51:37,480
is to say we shouldn't
make anybody worse off.

1162
00:51:37,480 --> 00:51:40,350
So if somebody reveals to us,
by their preference, which

1163
00:51:40,350 --> 00:51:42,660
is what these fellows are
doing here, that they really

1164
00:51:42,660 --> 00:51:44,370
don't want to receive
these letters,

1165
00:51:44,370 --> 00:51:48,120
we should respect that.

1166
00:51:48,120 --> 00:51:50,850
And then one potential
solution here

1167
00:51:50,850 --> 00:51:53,560
is to say, well, why don't
we just let people opt out,

1168
00:51:53,560 --> 00:51:55,810
where you just say, OK, you
send people these letters.

1169
00:51:55,810 --> 00:51:57,435
And then you have to
click on something

1170
00:51:57,435 --> 00:51:58,980
or you have to opt in or out.

1171
00:51:58,980 --> 00:52:00,630
And then maybe people
who really don't

1172
00:52:00,630 --> 00:52:05,070
want to receive these letters,
we can essentially just let

1173
00:52:05,070 --> 00:52:06,030
them be.

1174
00:52:06,030 --> 00:52:07,920
There's no judgment
here in terms

1175
00:52:07,920 --> 00:52:11,600
of is that socially
what we want to do

1176
00:52:11,600 --> 00:52:13,300
and is this a good
or a bad thing,

1177
00:52:13,300 --> 00:52:16,920
this is just to say we should
respect people's preferences

1178
00:52:16,920 --> 00:52:19,105
in certain ways.

1179
00:52:19,105 --> 00:52:19,980
Does that make sense?

1180
00:52:24,530 --> 00:52:27,260
Now the next question
you might ask

1181
00:52:27,260 --> 00:52:30,410
is, well, what form
should paternalism take?

1182
00:52:30,410 --> 00:52:32,690
And that's the thing,
since it's important,

1183
00:52:32,690 --> 00:52:36,980
because you could take
401(k) savings policies,

1184
00:52:36,980 --> 00:52:38,980
you can come up with a
lots of different options

1185
00:52:38,980 --> 00:52:43,380
with which you could design
your 401(k) savings policies.

1186
00:52:43,380 --> 00:52:45,510
You could have active decisions.

1187
00:52:45,510 --> 00:52:47,940
You can have defaults
for participation.

1188
00:52:47,940 --> 00:52:51,280
You have defaults for
all this information.

1189
00:52:51,280 --> 00:52:55,130
You have a personal financial
planner sitting down with you.

1190
00:52:55,130 --> 00:52:57,110
You might have a minimum
amount of-- this is,

1191
00:52:57,110 --> 00:52:58,860
I guess, hard paternalism--
minimum amount

1192
00:52:58,860 --> 00:53:01,710
of non-zero 401(k) savings
rate, which is just force people

1193
00:53:01,710 --> 00:53:02,730
to save.

1194
00:53:02,730 --> 00:53:06,210
And it's important to say it's
not just about paternalism

1195
00:53:06,210 --> 00:53:07,170
versus non-paternalism.

1196
00:53:07,170 --> 00:53:10,830
Once you get into the
paternalism realm,

1197
00:53:10,830 --> 00:53:12,570
it's actually very
hard to figure out

1198
00:53:12,570 --> 00:53:14,010
what should we do
because there's

1199
00:53:14,010 --> 00:53:15,093
so many different options.

1200
00:53:15,093 --> 00:53:19,100
And it's hard to tell which
one is better than the other.

1201
00:53:19,100 --> 00:53:21,500
More generally, even,
it's actually quite hard

1202
00:53:21,500 --> 00:53:24,680
to come up with any
non-paternalism option.

1203
00:53:24,680 --> 00:53:28,547
That is to say, think of all
the options that are there.

1204
00:53:28,547 --> 00:53:30,630
In some sense, once you
think that defaults affect

1205
00:53:30,630 --> 00:53:33,930
people's decisions,
then no savings default

1206
00:53:33,930 --> 00:53:35,280
is also a form of paternalism.

1207
00:53:35,280 --> 00:53:37,740
In some sense you say, you
choose something for people,

1208
00:53:37,740 --> 00:53:39,930
and that's going to
push them to not, say,

1209
00:53:39,930 --> 00:53:42,240
fully choose [INAUDIBLE]
like there's nothing special

1210
00:53:42,240 --> 00:53:47,340
about the zero [INAUDIBLE] also
have a 10% savings default.

1211
00:53:47,340 --> 00:53:50,040
And so it's important
to understand,

1212
00:53:50,040 --> 00:53:51,850
regardless of what
choice you choose,

1213
00:53:51,850 --> 00:53:54,660
you're going to affect people's
choices and their behaviors

1214
00:53:54,660 --> 00:53:56,020
in certain ways.

1215
00:53:56,020 --> 00:54:00,930
And so that's what it
means, essentially, that

1216
00:54:00,930 --> 00:54:07,080
have to take a stance in which
you affect people's behavior.

1217
00:54:07,080 --> 00:54:10,940
In a way, just saying we're just
choosing a zero savings default

1218
00:54:10,940 --> 00:54:13,350
is in some sense a cop-out.

1219
00:54:13,350 --> 00:54:15,750
Because even the
zero savings default

1220
00:54:15,750 --> 00:54:18,880
is affecting people's
behavior quite a bit.

1221
00:54:18,880 --> 00:54:22,320
So in a way, the discussion
should rather be about what is

1222
00:54:22,320 --> 00:54:26,610
the optimal way or the optimal
paternalism-- soft paternalism,

1223
00:54:26,610 --> 00:54:27,390
I would argue--

1224
00:54:27,390 --> 00:54:30,480
rather than saying
paternalism versus not.

1225
00:54:30,480 --> 00:54:32,685
Because regardless of
what option you choose,

1226
00:54:32,685 --> 00:54:34,950
you're going to affect
behaviors in certain ways

1227
00:54:34,950 --> 00:54:37,410
because precisely the way
you provide information

1228
00:54:37,410 --> 00:54:39,720
to people on their
default frames, et cetera,

1229
00:54:39,720 --> 00:54:42,640
affect people's choices.

1230
00:54:42,640 --> 00:54:43,560
Does that make sense?

1231
00:54:46,200 --> 00:54:47,040
OK.

1232
00:54:47,040 --> 00:54:51,690
So here's a very quick example
of a very useful and helpful

1233
00:54:51,690 --> 00:54:54,060
example of libertarian
paternalism

1234
00:54:54,060 --> 00:54:56,720
that has been quite successful.

1235
00:54:56,720 --> 00:55:00,270
So Thaler and
Benartzi were asked

1236
00:55:00,270 --> 00:55:04,410
to design a 401(k) savings for a
company that wanted to increase

1237
00:55:04,410 --> 00:55:06,570
its employees' savings.

1238
00:55:06,570 --> 00:55:09,060
So the company hired
a financial advisor.

1239
00:55:09,060 --> 00:55:11,070
The employees could sit
down with the advisor

1240
00:55:11,070 --> 00:55:13,650
to evaluate their financial
situation and make a plan.

1241
00:55:13,650 --> 00:55:17,040
So for those who agreed to talk
to the advisor, the advisor--

1242
00:55:17,040 --> 00:55:18,390
he, I guess, in this case--

1243
00:55:18,390 --> 00:55:19,860
recommended a savings rate.

1244
00:55:19,860 --> 00:55:21,270
Some agreed to implement this.

1245
00:55:21,270 --> 00:55:23,410
And then Benartzi and
Thaler got the leftovers.

1246
00:55:23,410 --> 00:55:26,790
So that's arguably a
group that's negatively

1247
00:55:26,790 --> 00:55:30,840
selected towards saving a lot.

1248
00:55:30,840 --> 00:55:34,680
Now, what's then the SMarT--
the Save More Tomorrow-- plan?

1249
00:55:34,680 --> 00:55:37,110
Employees were then
approached about increasing

1250
00:55:37,110 --> 00:55:40,110
their contribution rate
a considerable time

1251
00:55:40,110 --> 00:55:42,000
before the next
scheduled pay raise.

1252
00:55:42,000 --> 00:55:44,250
So that's to say, before the
next scheduled pay raise,

1253
00:55:44,250 --> 00:55:46,740
they were essentially asked
about whether they would

1254
00:55:46,740 --> 00:55:49,050
like to join
voluntarily and if they

1255
00:55:49,050 --> 00:55:53,010
wanted to do their
contribution to the plan

1256
00:55:53,010 --> 00:55:54,540
was going to be
increased beginning

1257
00:55:54,540 --> 00:55:57,120
with the first paycheck
after the raise.

1258
00:55:57,120 --> 00:55:59,350
So that's to say, I told
you, in December, you're

1259
00:55:59,350 --> 00:56:00,870
going to get a pay
raise, would you

1260
00:56:00,870 --> 00:56:04,020
like to increase your
contribution then

1261
00:56:04,020 --> 00:56:06,600
once you receive that pay raise?

1262
00:56:06,600 --> 00:56:11,700
And so then you could
choose a preset maximum.

1263
00:56:11,700 --> 00:56:14,850
And then what would happen
is the contribution rate

1264
00:56:14,850 --> 00:56:17,760
would automatically increase
over time with each pay raise

1265
00:56:17,760 --> 00:56:19,300
that you would receive.

1266
00:56:19,300 --> 00:56:22,560
Suppose your pay raise
is like, say, 5% or 3%,

1267
00:56:22,560 --> 00:56:24,240
what you would
get then is like--

1268
00:56:24,240 --> 00:56:25,475
and suppose it's 3%.

1269
00:56:25,475 --> 00:56:29,935
You would get only a 2% increase
in your pay and 1% or 1%

1270
00:56:29,935 --> 00:56:31,560
percentage point of
your pay would then

1271
00:56:31,560 --> 00:56:34,140
go towards the
retirement contribution.

1272
00:56:34,140 --> 00:56:35,970
Importantly, the
employee can opt out

1273
00:56:35,970 --> 00:56:38,610
of the plan at any point in
time by essentially just making

1274
00:56:38,610 --> 00:56:41,530
a phone call or the like.

1275
00:56:41,530 --> 00:56:46,540
Now, why is this plan a
libertarian paternalism?

1276
00:56:46,540 --> 00:56:49,430
Well because of aspects 2 and 4.

1277
00:56:49,430 --> 00:56:50,055
It's voluntary.

1278
00:56:50,055 --> 00:56:51,560
So you can do whatever you want.

1279
00:56:51,560 --> 00:56:54,010
You don't have to
choose this at all.

1280
00:56:54,010 --> 00:56:55,570
You can essentially decline it.

1281
00:56:55,570 --> 00:56:58,130
You can also opt out
at any point in time.

1282
00:56:58,130 --> 00:57:01,242
So arguably nobody
is made worse off.

1283
00:57:01,242 --> 00:57:02,950
Because essentially
what we're doing here

1284
00:57:02,950 --> 00:57:06,440
is increasing people's
options by saying,

1285
00:57:06,440 --> 00:57:08,990
here's an option for you
to increase your saving

1286
00:57:08,990 --> 00:57:11,150
automatically over time.

1287
00:57:11,150 --> 00:57:14,660
And you can leave it whenever
you would like to do.

1288
00:57:14,660 --> 00:57:18,200
So arguably nobody should
be worse off by this.

1289
00:57:18,200 --> 00:57:22,380
And potentially some people are
made better off quite a bit.

1290
00:57:22,380 --> 00:57:23,990
And that's exactly
what you see here.

1291
00:57:23,990 --> 00:57:26,930
Essentially, when you
look at the people who

1292
00:57:26,930 --> 00:57:28,860
joined the SMarT plan--

1293
00:57:28,860 --> 00:57:31,690
notice that this
is not random here.

1294
00:57:31,690 --> 00:57:34,630
So there's some concerns
about an indication.

1295
00:57:34,630 --> 00:57:38,103
But setting that aside, if you
look at the people who, again,

1296
00:57:38,103 --> 00:57:39,520
arguably, are
negatively selected,

1297
00:57:39,520 --> 00:57:44,280
these are people who declined
the financial consultant,

1298
00:57:44,280 --> 00:57:46,970
they essentially increased
their savings dramatically

1299
00:57:46,970 --> 00:57:49,970
as a consequence
of this SMarT plan.

1300
00:57:49,970 --> 00:57:53,150
And now the SMarT plan has
been introduced or implemented

1301
00:57:53,150 --> 00:57:55,377
in quite a few places.

1302
00:57:55,377 --> 00:57:56,960
But anyway, that's
not the point here.

1303
00:57:56,960 --> 00:57:59,300
I want to say what's
good about the plan is,

1304
00:57:59,300 --> 00:58:03,110
well, it's a SMarT plan not just
because of the acronym but also

1305
00:58:03,110 --> 00:58:06,200
because it's using behavioral
principles or stuff

1306
00:58:06,200 --> 00:58:09,750
that you learned in class
quite a bit and carefully.

1307
00:58:09,750 --> 00:58:11,195
So what are these principles?

1308
00:58:11,195 --> 00:58:13,760
So why is this a
useful plan to use?

1309
00:58:13,760 --> 00:58:17,210
And I'm going to show you
again what the plan is.

1310
00:58:17,210 --> 00:58:18,620
What's clever about this plan?

1311
00:58:18,620 --> 00:58:20,432
What is it using carefully?

1312
00:58:36,265 --> 00:58:38,890
By the way, I don't know if you
can hear this background noise.

1313
00:58:38,890 --> 00:58:42,910
My neighbor has engaged in
various home improvement

1314
00:58:42,910 --> 00:58:45,817
projects since the quarantine.

1315
00:58:48,499 --> 00:58:53,175
I have so far refrained
from complaining about it

1316
00:58:53,175 --> 00:58:55,300
because I don't know what's
going on in their lives

1317
00:58:55,300 --> 00:58:55,800
otherwise.

1318
00:58:55,800 --> 00:58:59,680
And maybe that's a way
of dealing with stress

1319
00:58:59,680 --> 00:59:02,670
and so on, if you can hear this.

1320
00:59:02,670 --> 00:59:07,620
Once you opt in, the default
is for the contribution rate

1321
00:59:07,620 --> 00:59:08,560
to increase.

1322
00:59:08,560 --> 00:59:11,430
So the default is essentially
working towards increasing

1323
00:59:11,430 --> 00:59:12,510
people's safety.

1324
00:59:12,510 --> 00:59:14,018
So there's two
things going on here.

1325
00:59:14,018 --> 00:59:16,560
One is present bias, when you
say, well, yeah, in the future,

1326
00:59:16,560 --> 00:59:17,693
I want to save more.

1327
00:59:17,693 --> 00:59:19,110
So precisely that's
kind of great.

1328
00:59:19,110 --> 00:59:22,690
Because now the future self
will be virtuous and so on.

1329
00:59:22,690 --> 00:59:25,000
So if people have present
bias, that will help.

1330
00:59:25,000 --> 00:59:27,660
And then there's
also the part that's

1331
00:59:27,660 --> 00:59:30,060
about paycheck increases.

1332
00:59:30,060 --> 00:59:32,190
Remember, people tend
to be loss-averse.

1333
00:59:32,190 --> 00:59:35,460
So what this plan is
also accomplishing is,

1334
00:59:35,460 --> 00:59:37,230
so if I'm asking you
right now, today,

1335
00:59:37,230 --> 00:59:39,850
would you like to increase your
savings rate, well, not only

1336
00:59:39,850 --> 00:59:41,490
are you present-biased,
but also you're

1337
00:59:41,490 --> 00:59:43,180
going to have a cut
in your paycheck.

1338
00:59:43,180 --> 00:59:45,315
And people hate cuts
in their paychecks.

1339
00:59:45,315 --> 00:59:47,190
So instead, it's saying,
well, in the future,

1340
00:59:47,190 --> 00:59:49,470
your raise will be lower.

1341
00:59:49,470 --> 00:59:54,300
And people are less averse
to reducing those increases

1342
00:59:54,300 --> 00:59:57,057
compared to
experiencing pay cuts.

1343
00:59:57,057 --> 00:59:58,640
So that's, I think,
what we have here.

1344
00:59:58,640 --> 01:00:00,750
So this essentially
present bias is helping.

1345
01:00:00,750 --> 01:00:02,580
Then there's
reference dependence.

1346
01:00:02,580 --> 01:00:05,475
And then there's defaults
that help people switching out

1347
01:00:05,475 --> 01:00:08,190
of some form of present bias,
where precisely as we said

1348
01:00:08,190 --> 01:00:09,990
previously, people
are not switching out

1349
01:00:09,990 --> 01:00:11,290
of those kinds of plans.

1350
01:00:11,290 --> 01:00:13,172
So if you think
that's a good thing

1351
01:00:13,172 --> 01:00:14,880
to increase people's
savings, here that's

1352
01:00:14,880 --> 01:00:16,410
[INAUDIBLE],, because
people might actually

1353
01:00:16,410 --> 01:00:17,970
procrastinate
switching out of it

1354
01:00:17,970 --> 01:00:19,990
once they're in in
the first place.

1355
01:00:19,990 --> 01:00:22,680
So that seems to
be a great example.

1356
01:00:22,680 --> 01:00:29,830
Because not only is it
preserving people's free choice

1357
01:00:29,830 --> 01:00:33,640
and not making people
worse off, but also it's

1358
01:00:33,640 --> 01:00:38,008
using, in a smart way,
behavioral principles.

1359
01:00:38,008 --> 01:00:39,550
Let me briefly talk
about the market.

1360
01:00:39,550 --> 01:00:41,925
And we talked about this a
little bit already previously,

1361
01:00:41,925 --> 01:00:43,810
but I want to make
this explicit.

1362
01:00:43,810 --> 01:00:48,130
So one argument that often
comes in particular out

1363
01:00:48,130 --> 01:00:52,677
of [INAUDIBLE] is,
well, if people have

1364
01:00:52,677 --> 01:00:54,510
self-control problems
or other issues, well,

1365
01:00:54,510 --> 01:00:57,170
will not the market come up with
some goods or some solutions

1366
01:00:57,170 --> 01:00:58,430
that will help people?

1367
01:00:58,430 --> 01:01:00,950
After all, there's
money to be made

1368
01:01:00,950 --> 01:01:02,690
in improving people's choices.

1369
01:01:02,690 --> 01:01:04,880
So if, really, self-control
problems or the like

1370
01:01:04,880 --> 01:01:06,770
are really important,
why is not anybody

1371
01:01:06,770 --> 01:01:09,410
coming up with a market solution
and a great product that

1372
01:01:09,410 --> 01:01:09,980
helps people?

1373
01:01:09,980 --> 01:01:13,220
And then they're going to make
a lot of money from doing so.

1374
01:01:13,220 --> 01:01:16,250
And so particularly if
consumers are sophisticated

1375
01:01:16,250 --> 01:01:18,830
and they demand ways to
change the consumption,

1376
01:01:18,830 --> 01:01:21,980
and then there will be somebody
who is providing that good.

1377
01:01:21,980 --> 01:01:24,410
And if the welfare
effects of those kinds

1378
01:01:24,410 --> 01:01:28,735
of self-control or other
distortions are large,

1379
01:01:28,735 --> 01:01:30,610
then you can make a lot
of money from helping

1380
01:01:30,610 --> 01:01:32,920
people improve their behavior.

1381
01:01:32,920 --> 01:01:38,990
So what's problematic with
this line of argument?

1382
01:01:38,990 --> 01:01:41,625
Well, the market is precisely
part of the problem.

1383
01:01:41,625 --> 01:01:43,250
So on the one hand,
you can say, if you

1384
01:01:43,250 --> 01:01:45,950
try to-- so if you Google--
and this is quite a while ago,

1385
01:01:45,950 --> 01:01:50,010
but right now it's, in some
sense, more sophisticated ways

1386
01:01:50,010 --> 01:01:51,900
and there's lots of
targeting going on.

1387
01:01:51,900 --> 01:01:54,453
But if you Google impulsive
credit-card spending, what

1388
01:01:54,453 --> 01:01:56,120
you're going to get
is a bunch of advice

1389
01:01:56,120 --> 01:01:57,225
to help you with that.

1390
01:01:57,225 --> 01:01:59,600
But at the same time, you also
get a bunch of credit card

1391
01:01:59,600 --> 01:02:02,090
offers or you're going to
get targeted eventually

1392
01:02:02,090 --> 01:02:05,690
by companies that try to
precisely exploit you.

1393
01:02:05,690 --> 01:02:07,670
So on the one
hand, there will be

1394
01:02:07,670 --> 01:02:09,920
a side of the market who
will potentially help you,

1395
01:02:09,920 --> 01:02:11,810
but there's another
side of the market who

1396
01:02:11,810 --> 01:02:13,520
will want to exploit you.

1397
01:02:13,520 --> 01:02:15,740
And importantly, what
the market will do

1398
01:02:15,740 --> 01:02:18,470
is firms will try to
target naive people.

1399
01:02:18,470 --> 01:02:20,660
And you might remember
this from lectures 5 and 6.

1400
01:02:20,660 --> 01:02:22,550
But essentially
firms precisely try

1401
01:02:22,550 --> 01:02:26,750
to find people who are perhaps
less sophisticated financially

1402
01:02:26,750 --> 01:02:27,870
in other ways.

1403
01:02:27,870 --> 01:02:30,560
So the firms are precisely
trying to target naive people.

1404
01:02:30,560 --> 01:02:32,930
Because those are the
people you can exploit.

1405
01:02:32,930 --> 01:02:36,440
Moreover, the naive are
the ones who will think,

1406
01:02:36,440 --> 01:02:39,290
well, I don't have any
problems, so I don't actually

1407
01:02:39,290 --> 01:02:42,240
need any goods or any sort of
products that might help you.

1408
01:02:42,240 --> 01:02:44,150
So not only are
they being targeted

1409
01:02:44,150 --> 01:02:47,450
by people who try
to screw them over,

1410
01:02:47,450 --> 01:02:51,290
but also they actually
are reluctant to seek help

1411
01:02:51,290 --> 01:02:55,790
because precisely they're
naive and they think it's not

1412
01:02:55,790 --> 01:02:57,510
even necessary to seek help.

1413
01:02:57,510 --> 01:03:00,360
So even a firm who wants to
help has trouble making money

1414
01:03:00,360 --> 01:03:00,860
with this.

1415
01:03:00,860 --> 01:03:02,990
Because people,
if they're naive,

1416
01:03:02,990 --> 01:03:07,950
will not necessarily understand
that help is required.

1417
01:03:07,950 --> 01:03:11,030
And that's precisely the tension
that we see in many markets,

1418
01:03:11,030 --> 01:03:14,088
where there's some
firms who might want

1419
01:03:14,088 --> 01:03:16,130
to help and support-- in
this case, self-control,

1420
01:03:16,130 --> 01:03:18,740
but it's also true
for other issues.

1421
01:03:18,740 --> 01:03:21,110
But there's, on the
other side, firms

1422
01:03:21,110 --> 01:03:24,140
that try to break down the
self-control from consumers

1423
01:03:24,140 --> 01:03:28,010
precisely because they can make
a bunch of money from that.

1424
01:03:28,010 --> 01:03:32,730
And so now the second
part, "severely limits

1425
01:03:32,730 --> 01:03:34,680
the market's ability to
provide self-control,"

1426
01:03:34,680 --> 01:03:36,810
because it essentially
[INAUDIBLE]..

1427
01:03:36,810 --> 01:03:40,200
Now, the government of course
can try to address this.

1428
01:03:40,200 --> 01:03:42,220
And this is what behavioral
IO would be about.

1429
01:03:42,220 --> 01:03:44,310
Behavioral IO would
be about saying--

1430
01:03:44,310 --> 01:03:45,810
Industrial Organization,
which would

1431
01:03:45,810 --> 01:03:48,180
be there's behavioral
consumers, there's

1432
01:03:48,180 --> 01:03:50,460
firms who try to
exploit consumers.

1433
01:03:50,460 --> 01:03:53,130
Now the government
is trying to support,

1434
01:03:53,130 --> 01:03:56,520
or by some form of regulation,
trying to help customers,

1435
01:03:56,520 --> 01:04:00,180
in particular things like the
Consumer Financial Protection

1436
01:04:00,180 --> 01:04:04,260
Bureau and so on trying
to help consumers not

1437
01:04:04,260 --> 01:04:07,860
to get exploited by banning
advertising, by capping

1438
01:04:07,860 --> 01:04:09,680
interest rates, and
so on and so forth,

1439
01:04:09,680 --> 01:04:14,590
but also by, for example,
mandating disclosure

1440
01:04:14,590 --> 01:04:17,640
and so on and so forth.

1441
01:04:17,640 --> 01:04:22,640
Now I'm going to say one final
thing about nudges, which is,

1442
01:04:22,640 --> 01:04:24,363
on the one hand, we
talk about nudges

1443
01:04:24,363 --> 01:04:26,030
in a sense of saying
that, look, there's

1444
01:04:26,030 --> 01:04:28,560
some behavior that
we're trying to improve,

1445
01:04:28,560 --> 01:04:31,040
which would be like suppose
people keep forgetting

1446
01:04:31,040 --> 01:04:32,720
certain things, to
take their medication

1447
01:04:32,720 --> 01:04:34,070
or to appear in court.

1448
01:04:34,070 --> 01:04:39,860
Let's send reminders
to improve behaviors.

1449
01:04:39,860 --> 01:04:42,920
We can set defaults in certain
ways and so on and so forth,

1450
01:04:42,920 --> 01:04:47,540
which is policies to improve
good behaviors, assuming

1451
01:04:47,540 --> 01:04:49,850
that the person who
makes those choices

1452
01:04:49,850 --> 01:04:52,980
is benevolent in certain ways.

1453
01:04:52,980 --> 01:04:54,920
Now, on the other
hand, of course,

1454
01:04:54,920 --> 01:04:57,080
firms can use the
same techniques

1455
01:04:57,080 --> 01:05:00,810
on the same approaches
to do the exact opposite

1456
01:05:00,810 --> 01:05:01,790
and so to harm people.

1457
01:05:01,790 --> 01:05:06,470
And this is what Thaler and
Sunstein called sludges, which

1458
01:05:06,470 --> 01:05:10,430
essentially is like using
these nudges in bad ways

1459
01:05:10,430 --> 01:05:14,520
by trying to push people
in wrong directions.

1460
01:05:14,520 --> 01:05:17,450
So an example of that would be
offering rebates on a product

1461
01:05:17,450 --> 01:05:22,010
and then having customers
required to mail in forms

1462
01:05:22,010 --> 01:05:24,140
and to, like-- and you
may have encountered these

1463
01:05:24,140 --> 01:05:26,240
and wondered why I have
to mail in this form.

1464
01:05:26,240 --> 01:05:27,410
And this is kind of tedious.

1465
01:05:27,410 --> 01:05:29,504
And then you end
up not doing it.

1466
01:05:29,504 --> 01:05:32,570
[INAUDIBLE] want
you to think, oh,

1467
01:05:32,570 --> 01:05:35,600
I'm getting this great discount,
and then actually never send it

1468
01:05:35,600 --> 01:05:37,850
in because then they don't
have to actually pay you.

1469
01:05:37,850 --> 01:05:40,623
They can sell you the
good, but then they

1470
01:05:40,623 --> 01:05:43,040
don't have to give you the
discount at the end of the day.

1471
01:05:43,040 --> 01:05:46,520
So there's a bunch
of types of cases,

1472
01:05:46,520 --> 01:05:49,330
both on the private sector but
also from the public sector,

1473
01:05:49,330 --> 01:05:51,320
such as things as
voter registration.

1474
01:05:51,320 --> 01:05:53,420
When you think about it,
why is voter registration

1475
01:05:53,420 --> 01:05:54,915
harder in certain places?

1476
01:05:54,915 --> 01:05:57,290
Well, the reason why it's hard
is because certain parties

1477
01:05:57,290 --> 01:05:59,930
and groups do not want certain
types of groups of people

1478
01:05:59,930 --> 01:06:00,860
to vote.

1479
01:06:00,860 --> 01:06:03,140
And those are essentially
deliberate ways

1480
01:06:03,140 --> 01:06:06,740
of making it hard for people to
make choices that aren't really

1481
01:06:06,740 --> 01:06:08,990
good for them or for society.

1482
01:06:08,990 --> 01:06:12,710
And that's what Thaler and
Sunstein would call sludges.

1483
01:06:12,710 --> 01:06:14,360
And then here it's
important to realize

1484
01:06:14,360 --> 01:06:17,270
that it's, in fact,
identifying and eliminating

1485
01:06:17,270 --> 01:06:20,090
such judges, which essentially
are certain ways that

1486
01:06:20,090 --> 01:06:22,670
make it hard for people to make
good choices for themselves

1487
01:06:22,670 --> 01:06:25,580
and for society, can be
equally important or perhaps as

1488
01:06:25,580 --> 01:06:28,310
important as sort of
implementing nudges

1489
01:06:28,310 --> 01:06:30,640
to improve people's
behavior in certain ways.

1490
01:06:30,640 --> 01:06:32,830
So it's not just
about taking behavior

1491
01:06:32,830 --> 01:06:35,000
where people, on their own,
screw up in certain ways

1492
01:06:35,000 --> 01:06:38,270
and trying to improve them
and trying to help them,

1493
01:06:38,270 --> 01:06:43,460
but rather helping people being
misdirected in certain ways

1494
01:06:43,460 --> 01:06:46,640
to be exploited.

1495
01:06:46,640 --> 01:06:48,950
And in many cases, in
particular in companies,

1496
01:06:48,950 --> 01:06:54,122
identify those kinds of sludges
and calling companies out

1497
01:06:54,122 --> 01:06:55,580
gets you pretty
far, in some sense,

1498
01:06:55,580 --> 01:06:57,320
on social media and the like.

1499
01:06:57,320 --> 01:06:59,450
Once called out,
companies often back down

1500
01:06:59,450 --> 01:07:01,400
and try to reduce
them because they

1501
01:07:01,400 --> 01:07:04,280
don't want to have bad
reputations of exploiting

1502
01:07:04,280 --> 01:07:05,600
customers.

1503
01:07:05,600 --> 01:07:08,520
OK, so let me sort of summarize
the lecture and then briefly,

1504
01:07:08,520 --> 01:07:10,670
in two minutes, the whole class.

1505
01:07:10,670 --> 01:07:13,940
So some forms of, in
particular, soft paternalism

1506
01:07:13,940 --> 01:07:15,470
can unambiguously
improve welfare.

1507
01:07:15,470 --> 01:07:17,220
And we should be quite
excited about that.

1508
01:07:17,220 --> 01:07:18,530
And they can help quite a bit.

1509
01:07:18,530 --> 01:07:21,110
Now, on the other hand, nudges
can also make things worse.

1510
01:07:21,110 --> 01:07:22,527
And particularly
if the government

1511
01:07:22,527 --> 01:07:25,040
doesn't know what's
good for people and we

1512
01:07:25,040 --> 01:07:26,870
want to be careful in nudging.

1513
01:07:26,870 --> 01:07:29,780
And there's some examples
[INAUDIBLE] in class.

1514
01:07:29,780 --> 01:07:33,640
But there's also some examples
of nudges actually backfiring.

1515
01:07:33,640 --> 01:07:36,330
Now, in addition to that, some
people dislike being nudged.

1516
01:07:36,330 --> 01:07:38,710
And we should respect that
and take that into account.

1517
01:07:38,710 --> 01:07:41,950
We try to maximize utility,
remember, as opposed

1518
01:07:41,950 --> 01:07:46,690
to certain behaviors or
take-up of certain outcomes.

1519
01:07:46,690 --> 01:07:49,030
And you want to be careful
not to make people worse off.

1520
01:07:49,030 --> 01:07:51,670
And helping people opt
out of certain nudges

1521
01:07:51,670 --> 01:07:54,670
or certain [? interventions ?]
should be quite helpful.

1522
01:07:54,670 --> 01:07:57,650
In addition, the market
does not solve everything.

1523
01:07:57,650 --> 01:08:00,310
So we should not have faith
in the market solving things.

1524
01:08:00,310 --> 01:08:04,340
In particular, the market
might make things worse.

1525
01:08:04,340 --> 01:08:08,170
And finally, reducing
sludges can go a long way

1526
01:08:08,170 --> 01:08:11,050
in improving behavior overall.

1527
01:08:11,050 --> 01:08:14,150
Now, the big-picture
summary of this class,

1528
01:08:14,150 --> 01:08:17,660
what is it, in one slide, that
I want you to take away and have

1529
01:08:17,660 --> 01:08:18,160
learned?

1530
01:08:18,160 --> 01:08:19,785
Of course, things
are more complicated.

1531
01:08:19,785 --> 01:08:22,370
But let me give you
a very brief summary.

1532
01:08:22,370 --> 01:08:24,744
So one, psychological
considerations

1533
01:08:24,744 --> 01:08:27,032
can be quite important in
a lot of economic choices

1534
01:08:27,032 --> 01:08:27,615
and behaviors.

1535
01:08:27,615 --> 01:08:29,859
And we should take
them very seriously.

1536
01:08:29,859 --> 01:08:32,109
For you, personally, I think
that means, in some ways,

1537
01:08:32,109 --> 01:08:35,020
understanding your own biases
and mistakes and issues better

1538
01:08:35,020 --> 01:08:37,270
and try to improve
decision-making.

1539
01:08:37,270 --> 01:08:38,810
It seems quite important.

1540
01:08:38,810 --> 01:08:41,750
That's true for short-run or
simple sort of considerations,

1541
01:08:41,750 --> 01:08:43,510
such as how to deal
with procrastination

1542
01:08:43,510 --> 01:08:45,160
and exercising and so on.

1543
01:08:45,160 --> 01:08:47,957
It's also true for long-run
considerations like which

1544
01:08:47,957 --> 01:08:50,290
jobs you want to have, what
friends do you want to keep,

1545
01:08:50,290 --> 01:08:51,457
and so on and so forth.

1546
01:08:51,457 --> 01:08:52,874
It could make you
[? both happy ?]

1547
01:08:52,874 --> 01:08:55,000
or unhappy in the future.

1548
01:08:55,000 --> 01:08:59,260
And you want to be quite
mindful about that.

1549
01:08:59,260 --> 01:09:02,859
Small changes or thinking
about things for a while

1550
01:09:02,859 --> 01:09:04,370
can make big differences.

1551
01:09:04,370 --> 01:09:08,170
So I very much encourage
you to introspect,

1552
01:09:08,170 --> 01:09:10,960
be more mindful or
mindful in your choices.

1553
01:09:10,960 --> 01:09:13,990
As I said last time,
experimenting, seeking advice,

1554
01:09:13,990 --> 01:09:15,970
including psychological
support and so on,

1555
01:09:15,970 --> 01:09:17,823
seem to be quite helpful.

1556
01:09:17,823 --> 01:09:19,990
At the same time, you also
shouldn't stress too much

1557
01:09:19,990 --> 01:09:25,210
and just try to be happy and
don't be too harsh on yourself.

1558
01:09:25,210 --> 01:09:27,817
Often, people are doing
quite well already overall.

1559
01:09:27,817 --> 01:09:29,859
It's also worth understanding
that sometimes it's

1560
01:09:29,859 --> 01:09:31,359
actually easier to
help your friends

1561
01:09:31,359 --> 01:09:33,020
or others than yourself.

1562
01:09:33,020 --> 01:09:34,810
It's often easier
to see, in others,

1563
01:09:34,810 --> 01:09:37,960
how they're misoptimizing and
how to improve their behaviors.

1564
01:09:37,960 --> 01:09:40,359
So in some ways,
helping others might

1565
01:09:40,359 --> 01:09:43,390
be, in fact one of the things
you might have taken away

1566
01:09:43,390 --> 01:09:44,979
from this class.

1567
01:09:44,979 --> 01:09:46,840
And then, finally, you
want to think about,

1568
01:09:46,840 --> 01:09:49,060
in particular, when you
think about designs of teams,

1569
01:09:49,060 --> 01:09:50,979
working with others,
incentives, products,

1570
01:09:50,979 --> 01:09:53,470
and so on, if you try to
design an app or any website

1571
01:09:53,470 --> 01:09:57,010
or [INAUDIBLE],, thinking about
psychological considerations

1572
01:09:57,010 --> 01:09:59,830
and designing things
[INAUDIBLE] is quite important

1573
01:09:59,830 --> 01:10:05,220
that you should [INAUDIBLE]
take those things into account.

1574
01:10:05,220 --> 01:10:06,880
Somebody is asking
about books to read.

1575
01:10:06,880 --> 01:10:08,443
So I just added them here.

1576
01:10:08,443 --> 01:10:10,110
I was planning to say
a little bit more,

1577
01:10:10,110 --> 01:10:12,840
but I'm out of time so I'm
just leaving this here.

1578
01:10:12,840 --> 01:10:15,900
I'll try to add a couple of
sentences for each [INAUDIBLE]..

1579
01:10:15,900 --> 01:10:19,770
There's some quite amazing
books that you can read.

1580
01:10:19,770 --> 01:10:22,940
In particular, for example,
Cialdini's '93 Influence book

1581
01:10:22,940 --> 01:10:24,800
is an amazing book to read.

1582
01:10:24,800 --> 01:10:26,760
Then Ariely has
really nice books,

1583
01:10:26,760 --> 01:10:28,280
including Predictably
Irrational.

1584
01:10:28,280 --> 01:10:31,515
And some of these are
more basic than others.

1585
01:10:31,515 --> 01:10:33,350
But if you're
interested in poverty,

1586
01:10:33,350 --> 01:10:35,060
Mullainathan and
Shafir's Scarcity book

1587
01:10:35,060 --> 01:10:37,880
is really an amazing read.

1588
01:10:37,880 --> 01:10:39,710
And then Danny
Kahneman's Thinking,

1589
01:10:39,710 --> 01:10:45,328
Fast and Slow is also extremely
insightful and interesting.

1590
01:10:45,328 --> 01:10:47,870
Of course there's other books,
for example, Dick Thaler talks

1591
01:10:47,870 --> 01:10:50,390
about the history of
behavioral economics,

1592
01:10:50,390 --> 01:10:53,960
Lewis is what's called
The Undoing Project, which

1593
01:10:53,960 --> 01:10:56,210
is about Kahneman,
first his friendship,

1594
01:10:56,210 --> 01:10:58,670
and how they sort of
essentially came to,

1595
01:10:58,670 --> 01:11:02,660
in some really important
ways, revolutionizing

1596
01:11:02,660 --> 01:11:03,890
some parts of economics.

1597
01:11:03,890 --> 01:11:05,810
So there's lots
of things to read.

1598
01:11:05,810 --> 01:11:08,102
If you're interested in any
of these, just let me know.

1599
01:11:08,102 --> 01:11:10,310
I'm very happy to chat.

1600
01:11:10,310 --> 01:11:12,380
So that's pretty much
all I have to say.

1601
01:11:12,380 --> 01:11:14,330
I had lots of fun
teaching this class.

1602
01:11:14,330 --> 01:11:20,180
I know this semester was tricky
and difficult for many of us.

1603
01:11:20,180 --> 01:11:22,890
Every time I teach this class,
I'm surprised how much I learn.

1604
01:11:22,890 --> 01:11:25,580
I hope you learned
something useful things too.

1605
01:11:25,580 --> 01:11:27,240
I guess this is from last time.

1606
01:11:27,240 --> 01:11:29,330
So there's no course
evaluations this year.

1607
01:11:29,330 --> 01:11:32,660
We might do an informal
one to just [INAUDIBLE]

1608
01:11:32,660 --> 01:11:35,330
try to learn more
about what you thought.

1609
01:11:35,330 --> 01:11:38,750
We did already an
informal evaluation,

1610
01:11:38,750 --> 01:11:41,320
but that's assuming that there
would be actual evaluations

1611
01:11:41,320 --> 01:11:42,360
this semester.

1612
01:11:42,360 --> 01:11:44,602
Unfortunately, MIT has
canceled them this year,

1613
01:11:44,602 --> 01:11:46,310
I think to be mindful
for some people who

1614
01:11:46,310 --> 01:11:50,810
are particularly affected
by the quarantine and so on.

1615
01:11:50,810 --> 01:11:54,010
Come to my office hours
if you have questions.

1616
01:11:54,010 --> 01:11:55,860
I have some Europe
opportunities.

1617
01:11:55,860 --> 01:11:58,190
Some of you have already
applied for those.

1618
01:11:58,190 --> 01:12:00,210
There's been quite
a bit of interest.

1619
01:12:00,210 --> 01:12:03,920
So if you have already applied,
I'll get back to you soon.

1620
01:12:03,920 --> 01:12:07,095
But any case, come and talk
to me if you're interested.

1621
01:12:07,095 --> 01:12:09,470
I hope you learn more economics
in the future [INAUDIBLE]

1622
01:12:09,470 --> 01:12:11,300
help you take psychological
considerations

1623
01:12:11,300 --> 01:12:14,240
in those economic
issues quite seriously.

1624
01:12:14,240 --> 01:12:16,390
So thank you very much.