1
00:00:00,040 --> 00:00:02,460
The following content is
provided under a Creative

2
00:00:02,460 --> 00:00:03,870
Commons license.

3
00:00:03,870 --> 00:00:06,910
Your support will help MIT
OpenCourseWare continue to

4
00:00:06,910 --> 00:00:10,550
offer high quality educational
resources for free.

5
00:00:10,550 --> 00:00:13,460
To make a donation or view
additional materials from

6
00:00:13,460 --> 00:00:17,390
hundreds of MIT courses, visit
MIT OpenCourseWare at

7
00:00:17,390 --> 00:00:18,640
ocw.mit.edu.

8
00:00:23,170 --> 00:00:26,510
PROFESSOR: We began our
discussion by talking about

9
00:00:26,510 --> 00:00:29,550
how society would choose
across different income

10
00:00:29,550 --> 00:00:36,180
distributions, which involve
the trade-off between

11
00:00:36,180 --> 00:00:39,750
preferences for more equal
distribution of income and the

12
00:00:39,750 --> 00:00:41,340
efficiency cost of transferring
income through

13
00:00:41,340 --> 00:00:45,300
both taxation and welfare
programs. We then moved on to

14
00:00:45,300 --> 00:00:47,680
talk about what the government
does in practice in terms of

15
00:00:47,680 --> 00:00:52,020
taxation and income
redistribution.

16
00:00:52,020 --> 00:00:54,390
What we're going to do today is
actually recognize that, in

17
00:00:54,390 --> 00:00:59,780
fact, most of government
spending, most of what we

18
00:00:59,780 --> 00:01:02,160
raise taxes for is not pure
income redistribution.

19
00:01:02,160 --> 00:01:04,780
We talked about things like the
TANF program, which gives

20
00:01:04,780 --> 00:01:07,180
money to single mothers, or the
SSI program, which gives

21
00:01:07,180 --> 00:01:08,850
money to disabled people.

22
00:01:08,850 --> 00:01:12,980
That's not most of what the
government raises money for.

23
00:01:12,980 --> 00:01:17,830
In fact, 95% of the transfers
that the government makes are

24
00:01:17,830 --> 00:01:19,155
not income based transfers.

25
00:01:19,155 --> 00:01:23,630
They're transfers of what we
call social insurance.

26
00:01:27,910 --> 00:01:31,166
So most of what the government
does today is not take from

27
00:01:31,166 --> 00:01:32,620
the rich and give to the poor.

28
00:01:32,620 --> 00:01:34,670
Most of what the government
does today is take from

29
00:01:34,670 --> 00:01:39,160
everyone to provide insurance
for everyone.

30
00:01:39,160 --> 00:01:44,990
Now basically, we talked in the
uncertainty lecture about

31
00:01:44,990 --> 00:01:47,960
why people would value
insurance.

32
00:01:47,960 --> 00:01:50,170
So essentially we think about
people choosing across states

33
00:01:50,170 --> 00:01:53,780
of nature like they're choosing
across goods.

34
00:01:53,780 --> 00:01:58,770
Then they will want to optimize
across those states,

35
00:01:58,770 --> 00:02:01,870
hit by a car versus not hit by
a car, the way they optimize

36
00:02:01,870 --> 00:02:04,200
across goods like
CDs and movies.

37
00:02:04,200 --> 00:02:06,430
And we said the way they
optimize across states is by

38
00:02:06,430 --> 00:02:07,340
buying insurance.

39
00:02:07,340 --> 00:02:09,750
That's the way you essentially
transfer resources from the

40
00:02:09,750 --> 00:02:11,230
state where something goes
bad to the state

41
00:02:11,230 --> 00:02:11,880
where something goes--

42
00:02:11,880 --> 00:02:14,300
I'm sorry, from the state where
something goes good to

43
00:02:14,300 --> 00:02:15,260
the state where it goes bad.

44
00:02:15,260 --> 00:02:17,740
The way you effect that
transfer, effectively, is by

45
00:02:17,740 --> 00:02:20,660
paying insurers when things go
good and getting from insurers

46
00:02:20,660 --> 00:02:21,960
when things go bad.

47
00:02:21,960 --> 00:02:25,340
That's insurance, and insurance
is an enormous share

48
00:02:25,340 --> 00:02:26,230
of the US economy.

49
00:02:26,230 --> 00:02:30,200
Currently, US as a whole spends
about $2 trillion a

50
00:02:30,200 --> 00:02:33,050
year on insuring ourselves
against adverse risks of

51
00:02:33,050 --> 00:02:34,250
various types.

52
00:02:34,250 --> 00:02:38,340
There's health insurance, which
is money we pay up front

53
00:02:38,340 --> 00:02:41,280
to insurers, and they cover our
medical costs should we

54
00:02:41,280 --> 00:02:43,290
get sick or become injured.

55
00:02:43,290 --> 00:02:48,160
There's disability insurance,
which is money that you pay in

56
00:02:48,160 --> 00:02:51,570
the form of taxation, and the
government uses that money to

57
00:02:51,570 --> 00:02:54,660
give you a benefit should
you become disabled.

58
00:02:54,660 --> 00:02:56,420
There's unemployment
insurance.

59
00:02:56,420 --> 00:02:57,180
We know about that.

60
00:02:57,180 --> 00:03:00,990
Unemployment insurance is money
that your employer pays

61
00:03:00,990 --> 00:03:04,230
in taxes, and in return, should
you lose your job due

62
00:03:04,230 --> 00:03:08,430
to a layoff you receive income
when you're laid off.

63
00:03:08,430 --> 00:03:10,370
That's certainly a topic
of much conversation

64
00:03:10,370 --> 00:03:12,540
these days, et cetera.

65
00:03:12,540 --> 00:03:15,340
So there are a number of social
insurance programs that

66
00:03:15,340 --> 00:03:18,990
are out there to help
individuals with the adverse

67
00:03:18,990 --> 00:03:20,510
risks they face.

68
00:03:20,510 --> 00:03:23,080
The big question we have to ask
in starting this lecture

69
00:03:23,080 --> 00:03:27,090
is, well, why does social
insurance exist?

70
00:03:27,090 --> 00:03:30,070
After all, there's actually
an enormous robust private

71
00:03:30,070 --> 00:03:30,890
insurance market.

72
00:03:30,890 --> 00:03:34,320
There's health insurance,
there's life insurance,

73
00:03:34,320 --> 00:03:38,170
there's casualty insurance.

74
00:03:38,170 --> 00:03:41,600
The private market insures
us for lots of risks.

75
00:03:41,600 --> 00:03:44,750
The question is why does the
government come in and also

76
00:03:44,750 --> 00:03:46,700
provide us this social
insurance?

77
00:03:46,700 --> 00:03:51,855
So we individuals will demand
insurance for risk.

78
00:03:51,855 --> 00:03:55,120
We know private insurance exists
for things like health

79
00:03:55,120 --> 00:03:57,070
insurance, life insurance, fire

80
00:03:57,070 --> 00:03:58,560
insurance, auto insurance.

81
00:03:58,560 --> 00:04:02,680
We know it exists, and that adds
up to about $2 trillion

82
00:04:02,680 --> 00:04:04,470
of private spending.

83
00:04:04,470 --> 00:04:07,540
Yet at the same time the
government also provides on

84
00:04:07,540 --> 00:04:10,380
the order of another trillion
dollars of social insurance

85
00:04:10,380 --> 00:04:13,250
spending every year for things
like health insurance, and

86
00:04:13,250 --> 00:04:15,120
disability, and other things.

87
00:04:15,120 --> 00:04:17,820
So the issue is why does the
government need to do that?

88
00:04:17,820 --> 00:04:19,430
Why does the government
need to roll

89
00:04:19,430 --> 00:04:20,769
in and provide insurance?

90
00:04:20,769 --> 00:04:22,280
We know individuals want it.

91
00:04:22,280 --> 00:04:25,230
We know a private market exists
that provides it.

92
00:04:25,230 --> 00:04:28,360
What's the market failure
that justifies a

93
00:04:28,360 --> 00:04:31,620
government role here?

94
00:04:31,620 --> 00:04:33,300
Think about the role of the
government in the economy--

95
00:04:33,300 --> 00:04:35,810
and this is a very small
part of this course--

96
00:04:35,810 --> 00:04:39,000
but the focus of the other
course I teach, 14.41.

97
00:04:39,000 --> 00:04:43,010
The basic question is in this
course we've largely learned

98
00:04:43,010 --> 00:04:45,640
the governments serves to screw
up the market if the

99
00:04:45,640 --> 00:04:47,090
market's working well.

100
00:04:47,090 --> 00:04:49,340
The only time the government
can help is if there's a

101
00:04:49,340 --> 00:04:51,350
market failure, if
there's something

102
00:04:51,350 --> 00:04:52,530
wrong with the market.

103
00:04:52,530 --> 00:04:56,090
Well in this case, what could
the market failure be?

104
00:04:56,090 --> 00:04:59,030
And the market failure, in the
context of insurance, is the

105
00:04:59,030 --> 00:05:00,660
problem of asymmetric
information.

106
00:05:05,280 --> 00:05:10,070
Asymmetric information,
information that is held

107
00:05:10,070 --> 00:05:12,250
differentially between parties
on either side of a

108
00:05:12,250 --> 00:05:13,500
transaction.

109
00:05:15,260 --> 00:05:17,860
And that when there's asymmetric
information, when

110
00:05:17,860 --> 00:05:19,910
different parties know different
amounts about a

111
00:05:19,910 --> 00:05:26,080
transaction, that can lead
to market failures.

112
00:05:26,080 --> 00:05:29,210
So the classic example of this
comes from the Nobel Prize

113
00:05:29,210 --> 00:05:31,120
winning economist
George Akerlof.

114
00:05:31,120 --> 00:05:33,990
It's called the lemons
problem.

115
00:05:33,990 --> 00:05:38,330
The lemons problem is,
basically, he did the example

116
00:05:38,330 --> 00:05:39,930
of the used car market.

117
00:05:39,930 --> 00:05:42,600
Imagine you've got a used car
market, which is a set of

118
00:05:42,600 --> 00:05:45,670
individuals who want to sell
their cars and a set of

119
00:05:45,670 --> 00:05:48,840
individuals who want
to buy those cars.

120
00:05:48,840 --> 00:05:51,220
And let's start with the full
information benchmark.

121
00:05:51,220 --> 00:05:54,490
Let's imagine there was perfect
information about the

122
00:05:54,490 --> 00:05:56,360
exact quality of every car.

123
00:05:56,360 --> 00:05:58,640
Imagine there was some test
you could do which would

124
00:05:58,640 --> 00:06:01,040
perfectly tell you exactly how
long the car is going to last,

125
00:06:01,040 --> 00:06:03,110
how well it's going
to run, et cetera.

126
00:06:03,110 --> 00:06:05,680
In that world there will be
a set of transactions.

127
00:06:05,680 --> 00:06:07,530
Some individuals will want to
sell their cars because they

128
00:06:07,530 --> 00:06:09,290
want to buy up, other individual
will be happy to

129
00:06:09,290 --> 00:06:12,980
pay that price, there will
be a set of transactions.

130
00:06:12,980 --> 00:06:16,340
Now imagine we go to the real
world, especially a real world

131
00:06:16,340 --> 00:06:19,065
when Akerlof wrote his article
in the '70s when, basically,

132
00:06:19,065 --> 00:06:21,210
if you went to buy a used car
you had to look at it and

133
00:06:21,210 --> 00:06:23,470
decide does it look OK.

134
00:06:23,470 --> 00:06:25,420
There's clearly asymmetric
information.

135
00:06:25,420 --> 00:06:28,860
Clearly the seller of the car
knows much more about its

136
00:06:28,860 --> 00:06:31,990
underlying qualities than
the buyer of the car.

137
00:06:31,990 --> 00:06:34,050
And in particular, the buyer
should be suspicious saying,

138
00:06:34,050 --> 00:06:36,820
wait a second, why are
you selling this car?

139
00:06:36,820 --> 00:06:40,250
If it's such a good car why
are you selling it?

140
00:06:40,250 --> 00:06:42,570
And the very fact that I see
you selling it makes me

141
00:06:42,570 --> 00:06:45,340
suspicious that something's
wrong with it.

142
00:06:45,340 --> 00:06:48,110
Given that I'm suspicious
something's wrong with it I'm

143
00:06:48,110 --> 00:06:50,930
going to be willing to pay less
for it than I would if I

144
00:06:50,930 --> 00:06:53,950
knew for sure it was OK.

145
00:06:53,950 --> 00:06:57,070
The fact that I don't have full
information as a buyer

146
00:06:57,070 --> 00:07:00,210
means that I'm going to pay less
for it than I would if I

147
00:07:00,210 --> 00:07:02,480
had full information.

148
00:07:02,480 --> 00:07:05,360
Well if I'm going to pay less
for it but the seller has a

149
00:07:05,360 --> 00:07:07,266
certain amount he's willing to
sell it at, the seller will

150
00:07:07,266 --> 00:07:08,680
say, well, I'm not going to
sell it to you for less.

151
00:07:08,680 --> 00:07:10,975
I have a certain supply curve
at which I want to sell.

152
00:07:14,520 --> 00:07:16,940
If I can get x for the car I'll
sell it, but if you're

153
00:07:16,940 --> 00:07:19,150
not going to pay me x I'm
not going to sell it.

154
00:07:19,150 --> 00:07:21,520
So the sale doesn't happen.

155
00:07:21,520 --> 00:07:24,240
So what you have is a car where
in a full information

156
00:07:24,240 --> 00:07:27,050
world a sale would happen, both
parties would be happy

157
00:07:27,050 --> 00:07:27,990
with that sale.

158
00:07:27,990 --> 00:07:29,830
But in an asymmetric information
world, since the

159
00:07:29,830 --> 00:07:33,640
first party, since the buyer's
suspicious of the quality and

160
00:07:33,640 --> 00:07:36,180
is therefore willing to pay
less, and the seller won't

161
00:07:36,180 --> 00:07:38,420
accept less the sale
doesn't happen.

162
00:07:38,420 --> 00:07:40,820
And that's a market failure.

163
00:07:40,820 --> 00:07:43,790
Relative to the 14.01 world we
started with, where everybody

164
00:07:43,790 --> 00:07:46,070
knows everything, where there's
full information, the

165
00:07:46,070 --> 00:07:48,330
market has failed because sales
which would make both

166
00:07:48,330 --> 00:07:51,230
parties better off do not
occur due to asymmetric

167
00:07:51,230 --> 00:07:52,480
information.

168
00:07:54,110 --> 00:07:57,690
So that is an example of the
market failure caused by

169
00:07:57,690 --> 00:08:00,640
asymmetric information.

170
00:08:00,640 --> 00:08:04,840
Now let's flip this around
and talk about insurance.

171
00:08:04,840 --> 00:08:07,760
In insurance it's the
opposite case.

172
00:08:07,760 --> 00:08:11,770
In insurance, I, the person
demanding the insurance, know

173
00:08:11,770 --> 00:08:14,255
more about me than the insurers

174
00:08:14,255 --> 00:08:16,330
who's selling me insurance.

175
00:08:16,330 --> 00:08:20,430
I know whether I'm someone who
gets sick a lot, whether I

176
00:08:20,430 --> 00:08:22,980
like skydiving, whether I
engage in other risky

177
00:08:22,980 --> 00:08:27,480
behaviors that may have health
problems for me, whether I am

178
00:08:27,480 --> 00:08:29,790
someone who smokes in bed so I'm
likely to catch my house

179
00:08:29,790 --> 00:08:32,890
on fire so that I really want
house insurance because I'm

180
00:08:32,890 --> 00:08:34,289
going to catch my house
on fire, et cetera.

181
00:08:34,289 --> 00:08:37,770
I know all that stuff about
me, the insurer doesn't.

182
00:08:37,770 --> 00:08:41,630
So the insurer, as a result,
might be unwilling to sell me

183
00:08:41,630 --> 00:08:44,650
insurance at a fair price.

184
00:08:44,650 --> 00:08:48,860
So to put this in numbers, let's
consider an example.

185
00:08:52,060 --> 00:08:57,370
Imagine that you are starting up
a health insurance program

186
00:08:57,370 --> 00:08:59,460
for MIT graduates.

187
00:08:59,460 --> 00:09:01,600
You want to say look, MIT
graduates are a pretty healthy

188
00:09:01,600 --> 00:09:05,520
lot, and so I'm going to start
up a health insurance program

189
00:09:05,520 --> 00:09:06,010
for MIT graduates.

190
00:09:06,010 --> 00:09:07,380
I'm going to make money since
they're basically pretty

191
00:09:07,380 --> 00:09:09,363
healthy, but they're
risk averse so they

192
00:09:09,363 --> 00:09:10,730
will buy the insurance.

193
00:09:10,730 --> 00:09:12,670
So imagine the facts
are the following.

194
00:09:12,670 --> 00:09:17,970
Imagine that for every for every
100 MIT graduates, over

195
00:09:17,970 --> 00:09:25,750
the next year, 90 of them will
spend nothing on medical care,

196
00:09:25,750 --> 00:09:31,140
and 10 of them will
spend $1,000.

197
00:09:31,140 --> 00:09:33,030
But we don't know which ones,
there's uncertainty.

198
00:09:33,030 --> 00:09:35,860
So I'm left with this
90% chance of zero,

199
00:09:35,860 --> 00:09:37,680
10% chance of $1,000.

200
00:09:37,680 --> 00:09:40,900
So everybody graduating, since
we're a pretty healthy crop,

201
00:09:40,900 --> 00:09:45,120
90% of you will spend nothing
but 10% of you will spend

202
00:09:45,120 --> 00:09:47,450
$1,000, and you don't
know which.

203
00:09:52,770 --> 00:09:53,920
And you're going to start
an insurance.

204
00:09:53,920 --> 00:09:55,820
You could say look, I'm going
to start an insurance.

205
00:09:55,820 --> 00:09:57,420
What I'm going to do is I'm
going to say is what's the

206
00:09:57,420 --> 00:10:01,210
expected payout that I'm going
to have as an insurer

207
00:10:01,210 --> 00:10:02,790
over the next year?

208
00:10:02,790 --> 00:10:07,440
Well, the expected payout over
the next year is $10,000

209
00:10:07,440 --> 00:10:08,970
total, right?

210
00:10:08,970 --> 00:10:10,780
And let's say there's
100 students you are

211
00:10:10,780 --> 00:10:12,030
going to sell to.

212
00:10:14,350 --> 00:10:18,690
The expected payout the next
year is $10,000 total.

213
00:10:18,690 --> 00:10:20,770
10 students will cost you $1,000
nine students will cost

214
00:10:20,770 --> 00:10:21,640
you nothing.

215
00:10:21,640 --> 00:10:25,800
Expected payout next year is
$10,000, or per student you're

216
00:10:25,800 --> 00:10:32,480
selling to the expected
payout is $100.

217
00:10:32,480 --> 00:10:32,740
Right?

218
00:10:32,740 --> 00:10:33,350
Have I got that right?

219
00:10:33,350 --> 00:10:33,830
Yeah.

220
00:10:33,830 --> 00:10:37,810
$100 per student you're
selling to.

221
00:10:37,810 --> 00:10:44,750
So if you sell to 100 students
at $100 you will break even.

222
00:10:44,750 --> 00:10:47,490
You will collect $10,000
at an expectation

223
00:10:47,490 --> 00:10:50,520
you'll pay out $10,000.

224
00:10:50,520 --> 00:10:58,080
So if you roll in and offer
insurance for $100 and you

225
00:10:58,080 --> 00:11:02,220
sell to all these students
then you'll break even.

226
00:11:02,220 --> 00:11:03,630
Let's say you then said
fine, I'm going to

227
00:11:03,630 --> 00:11:07,620
offer it for $110.

228
00:11:07,620 --> 00:11:12,320
Well, as long as people are
somewhat risk averse, in our

229
00:11:12,320 --> 00:11:16,690
standard model they'd still
like insurance for $110.

230
00:11:16,690 --> 00:11:19,160
They'd still be willing to pay
a little risk premium--

231
00:11:19,160 --> 00:11:21,600
remember about risk premiums
to get health insurance.

232
00:11:21,600 --> 00:11:25,515
So you still sell it, let's
say, and you make $1,000.

233
00:11:25,515 --> 00:11:27,500
Then you're going to sell to
100 students at 10 extra

234
00:11:27,500 --> 00:11:29,290
dollars to make $1,000 profit.

235
00:11:29,290 --> 00:11:30,860
So this is a pretty
good product.

236
00:11:30,860 --> 00:11:34,900
I'll make a $1,000
profit on this.

237
00:11:34,900 --> 00:11:39,390
Now, what's in fact the problem
with this analysis?

238
00:11:39,390 --> 00:11:42,340
The problem is that what
if some of the

239
00:11:42,340 --> 00:11:44,440
healthy guys say, look--

240
00:11:44,440 --> 00:11:47,860
what if people know whether
they're healthy or sick?

241
00:11:47,860 --> 00:11:49,120
What if it's not uncertain?

242
00:11:49,120 --> 00:11:51,450
Or what if people have
a good inkling?

243
00:11:51,450 --> 00:11:56,750
What if, in fact, half of the 90
people who will spend zero,

244
00:11:56,750 --> 00:11:59,650
what if 50 of them actually
know they're

245
00:11:59,650 --> 00:12:00,690
going to spend zero?

246
00:12:00,690 --> 00:12:01,700
What if, in fact, the
story is different?

247
00:12:01,700 --> 00:12:04,400
It's not totally uncertain,
but 50 people know they're

248
00:12:04,400 --> 00:12:04,810
going to spend zero.

249
00:12:04,810 --> 00:12:07,190
They're really healthy and they
don't like the doctor,

250
00:12:07,190 --> 00:12:09,210
they know they're going
to spend zero.

251
00:12:09,210 --> 00:12:12,550
Whereas the other 50, well, 40
of them might spend zero and

252
00:12:12,550 --> 00:12:14,210
10 might spend $1,000.

253
00:12:14,210 --> 00:12:16,445
Well in that case, the 50 that
knows they're going to spend

254
00:12:16,445 --> 00:12:18,410
zero won't buy the
health insurance.

255
00:12:18,410 --> 00:12:20,220
They'll be like why
should I buy it?

256
00:12:20,220 --> 00:12:22,235
I know I'm going to spend zero,
why should I pay $100 or

257
00:12:22,235 --> 00:12:26,230
$110 dollars for health
insurance?

258
00:12:26,230 --> 00:12:28,920
So then, how much
will you sell?

259
00:12:28,920 --> 00:12:33,010
And let's say you charge
$110 to make a profit.

260
00:12:33,010 --> 00:12:36,280
You'll sell to 40
guys, 40 healthy

261
00:12:36,280 --> 00:12:38,770
guys, and 10 sick guys.

262
00:12:38,770 --> 00:12:47,640
So 50 guys times $110, so
you'll make $5,500.

263
00:12:47,640 --> 00:12:49,010
What will you pay out?

264
00:12:49,010 --> 00:12:56,260
You'll pay out $10,000 because
the 10 sick guys are

265
00:12:56,260 --> 00:12:58,440
definitely buying, right?

266
00:12:58,440 --> 00:13:00,310
The 10 guys who'll end up
getting sick are definitely

267
00:13:00,310 --> 00:13:02,540
buying along with
40 healthy guys.

268
00:13:02,540 --> 00:13:05,710
So you will collect $5,500 but
you'll pay out $10,000.

269
00:13:05,710 --> 00:13:08,320
You'll lose money.

270
00:13:08,320 --> 00:13:09,710
Why will you lose money?

271
00:13:09,710 --> 00:13:15,930
You'll lose money because you'll
be collecting as if

272
00:13:15,930 --> 00:13:18,600
you're dealing with a full
probability distribution but

273
00:13:18,600 --> 00:13:20,360
you're only going to see
a truncated probability

274
00:13:20,360 --> 00:13:21,540
distribution.

275
00:13:21,540 --> 00:13:23,830
You're only going to see the
distribution for the people

276
00:13:23,830 --> 00:13:25,170
who have some chance
of getting sick.

277
00:13:25,170 --> 00:13:27,110
The very healthy people
drop out.

278
00:13:27,110 --> 00:13:30,790
As a result you lose money.

279
00:13:30,790 --> 00:13:32,680
This is the insurer problem.

280
00:13:32,680 --> 00:13:36,300
If the healthiest people don't
buy insurance then they can't

281
00:13:36,300 --> 00:13:38,350
price it fairly.

282
00:13:38,350 --> 00:13:39,430
How would you overcome this?

283
00:13:39,430 --> 00:13:43,010
Well, you would have to raise
your price a lot, right?

284
00:13:43,010 --> 00:13:45,170
So let's say you say fine, I'm
going to overcome this.

285
00:13:45,170 --> 00:13:50,850
If I raise my price to
$200 or to $210 then

286
00:13:50,850 --> 00:13:52,410
I'd make money again.

287
00:13:52,410 --> 00:13:55,140
If I raise my price
$210 that's 50

288
00:13:55,140 --> 00:13:56,560
people buying at $210.

289
00:13:56,560 --> 00:14:00,740
Then I get $11,000 in premiums
and I don't have to spend

290
00:14:00,740 --> 00:14:04,730
$10,000, and I make
money again.

291
00:14:04,730 --> 00:14:06,690
What's the problem with
that solution?

292
00:14:06,690 --> 00:14:07,050
Yeah?

293
00:14:07,050 --> 00:14:08,916
AUDIENCE: People will
probably be dropping

294
00:14:08,916 --> 00:14:09,950
out because you just--

295
00:14:09,950 --> 00:14:11,650
PROFESSOR: Some of these 40 will
say, well, I was happy to

296
00:14:11,650 --> 00:14:13,670
pay $100 but I'm not
happy to pay $200.

297
00:14:13,670 --> 00:14:15,720
The chances are really low
I'm going to get sick.

298
00:14:15,720 --> 00:14:16,750
So this goes down.

299
00:14:16,750 --> 00:14:18,430
So you collect less money
but you still pay

300
00:14:18,430 --> 00:14:19,590
out to the sick people.

301
00:14:19,590 --> 00:14:21,780
And you never win.

302
00:14:21,780 --> 00:14:22,470
You can never win.

303
00:14:22,470 --> 00:14:24,550
As you raise the price the
healthier and healthier people

304
00:14:24,550 --> 00:14:28,070
drop out until you just simply
can't make money.

305
00:14:28,070 --> 00:14:28,425
Yeah?

306
00:14:28,425 --> 00:14:31,425
AUDIENCE: Is there any situation
where if you have a

307
00:14:31,425 --> 00:14:34,360
large enough number of people
that you're supplying to you

308
00:14:34,360 --> 00:14:37,212
can set the price at a certain
level such that you're

309
00:14:37,212 --> 00:14:41,646
guaranteed that there are at
least going to be a large

310
00:14:41,646 --> 00:14:45,139
number of people left after
dropping out so that you won't

311
00:14:45,139 --> 00:14:47,070
lose money?

312
00:14:47,070 --> 00:14:48,610
PROFESSOR: Yes, and what's
that going to depend on?

313
00:14:48,610 --> 00:14:52,985
So one extreme example is what
I've described here, where as

314
00:14:52,985 --> 00:14:55,113
you raise the price more healthy
drop out and so you

315
00:14:55,113 --> 00:14:56,550
can never make money.

316
00:14:56,550 --> 00:15:00,060
Under what conditions will that
be a particular problem

317
00:15:00,060 --> 00:15:02,440
or not be a problem?

318
00:15:02,440 --> 00:15:04,920
Well there's two key factors.

319
00:15:04,920 --> 00:15:06,605
First of all, it's going
to be how imperfect

320
00:15:06,605 --> 00:15:08,450
the information is.

321
00:15:08,450 --> 00:15:11,880
For example, if people know for
sure whether they're going

322
00:15:11,880 --> 00:15:14,990
to be in the zero bucket or
the $1,000 bucket then by

323
00:15:14,990 --> 00:15:17,190
definition you can't
make money.

324
00:15:17,190 --> 00:15:19,960
If you know for sure, with 100%
certainty, then why would

325
00:15:19,960 --> 00:15:21,500
you ever buy insurance unless
you're going to be in the

326
00:15:21,500 --> 00:15:23,080
$1,000 bucket?

327
00:15:23,080 --> 00:15:26,100
And why would you ever pay
more than $1,000 for it?

328
00:15:26,100 --> 00:15:28,650
So there's no way to make money
if people know for sure.

329
00:15:28,650 --> 00:15:33,450
So the first thing is where
nobody has any idea at all.

330
00:15:33,450 --> 00:15:35,830
Then this is going to be less
of a problem because healthy

331
00:15:35,830 --> 00:15:38,210
people won't drop out because
nobody knows who's healthy.

332
00:15:38,210 --> 00:15:40,110
So the first issue is going to
be the asymmetry of your

333
00:15:40,110 --> 00:15:41,490
information.

334
00:15:41,490 --> 00:15:44,260
The second factors is
going to be what?

335
00:15:44,260 --> 00:15:47,590
The level of risk aversion,
because if people are very

336
00:15:47,590 --> 00:15:50,980
risk averse then they may be
willing to pay a lot for

337
00:15:50,980 --> 00:15:54,660
insurance even if the odds are
low it's going to hurt them.

338
00:15:54,660 --> 00:15:57,310
So in this case, once again if
people know for sure, risk

339
00:15:57,310 --> 00:15:58,390
aversion doesn't matter.

340
00:15:58,390 --> 00:16:00,610
But if people don't know for
sure, if there's some

341
00:16:00,610 --> 00:16:03,520
uncertainty, then you may get
healthy people buying health

342
00:16:03,520 --> 00:16:06,960
insurance because they're
risk averse.

343
00:16:06,960 --> 00:16:08,820
They may buy it anyway, and
so then you may not

344
00:16:08,820 --> 00:16:10,220
have a market failure.

345
00:16:10,220 --> 00:16:13,010
So basically, the more asymmetry
of information there

346
00:16:13,010 --> 00:16:15,580
is the more likely you're
going to lose money?

347
00:16:15,580 --> 00:16:17,060
What happens if you're
going to lose money?

348
00:16:17,060 --> 00:16:18,955
You're not going to start your
insurance company and MIT

349
00:16:18,955 --> 00:16:20,800
students will end
up uninsured.

350
00:16:20,800 --> 00:16:22,640
That's a market failure.

351
00:16:22,640 --> 00:16:26,130
The market for insurance will
fail, the market for insurance

352
00:16:26,130 --> 00:16:29,840
will fail because you will
end up not being

353
00:16:29,840 --> 00:16:31,680
able to make money.

354
00:16:31,680 --> 00:16:35,290
You're not going to be able to
make money because only the

355
00:16:35,290 --> 00:16:38,710
sickest people will buy, and
you lose by definition.

356
00:16:38,710 --> 00:16:42,070
That's the problem that we have
with private insurance.

357
00:16:42,070 --> 00:16:44,800
If the healthiest people stay
out, then the insurer can't

358
00:16:44,800 --> 00:16:46,220
make money.

359
00:16:46,220 --> 00:16:48,210
The insurers need a distribution
of healthy and

360
00:16:48,210 --> 00:16:50,290
sick to make money, they
can't make money if the

361
00:16:50,290 --> 00:16:51,540
healthy stay out.

362
00:16:54,100 --> 00:16:58,430
So what can we do about this?

363
00:16:58,430 --> 00:16:59,110
What can we do about this?

364
00:16:59,110 --> 00:17:02,800
Well there's two classes of
solutions we can have. The

365
00:17:02,800 --> 00:17:04,854
first is we can subsidize.

366
00:17:09,369 --> 00:17:16,540
Imagine if we said, we MIT, will
roll in and provide the

367
00:17:16,540 --> 00:17:18,040
insurance for free
to everyone.

368
00:17:20,660 --> 00:17:23,099
And then we'll pay the insurer
what they want.

369
00:17:23,099 --> 00:17:26,245
So MIT will come in, they'll
absorb the cost of the insurer

370
00:17:26,245 --> 00:17:28,270
and give it to MIT students
for free.

371
00:17:28,270 --> 00:17:31,135
Well, in that case all the MIT
students would take it, it

372
00:17:31,135 --> 00:17:35,430
would be free, and the insurer
could charge $100 or $110

373
00:17:35,430 --> 00:17:37,390
based on the negotiation
with MIT.

374
00:17:37,390 --> 00:17:39,625
And the insurer will provide the
insurance because everyone

375
00:17:39,625 --> 00:17:40,310
would take it.

376
00:17:40,310 --> 00:17:42,420
And MIT would just
pay the cost.

377
00:17:42,420 --> 00:17:44,670
So the market failure would be
solved, but of course at a

378
00:17:44,670 --> 00:17:47,720
cost of MIT having to pay
$10,000 or $11,000, or

379
00:17:47,720 --> 00:17:49,930
whatever it's going to be.

380
00:17:49,930 --> 00:17:54,250
So the market failure would be
solved by subsidizing the

381
00:17:54,250 --> 00:17:55,610
healthy people to come in.

382
00:17:55,610 --> 00:17:57,830
So essentially, if you pay off
the best risks to come in the

383
00:17:57,830 --> 00:18:00,010
market you solve the market
failure because then the

384
00:18:00,010 --> 00:18:03,620
insurer knows they're getting a
good distribution of risks.

385
00:18:03,620 --> 00:18:06,270
What's the other thing
you can do?

386
00:18:06,270 --> 00:18:07,920
Well the other thing you can do
is the centerpiece of the

387
00:18:07,920 --> 00:18:09,760
current health care reform bill,
which I'll talk about

388
00:18:09,760 --> 00:18:12,920
next time, which is you can
mandate that everyone buy

389
00:18:12,920 --> 00:18:14,570
health insurance.

390
00:18:14,570 --> 00:18:18,390
If there was a law that every
MIT student had to have health

391
00:18:18,390 --> 00:18:21,950
insurance when they graduate
then, once again, the insurer

392
00:18:21,950 --> 00:18:25,140
would know what the distribution
of risks are and

393
00:18:25,140 --> 00:18:27,780
they could sell insurance
at a fair price.

394
00:18:27,780 --> 00:18:28,832
They wouldn't have to worry if
the healthy drop out because

395
00:18:28,832 --> 00:18:30,082
the healthy couldn't,
it would be illegal.

396
00:18:33,570 --> 00:18:35,060
Now, these two solutions--

397
00:18:35,060 --> 00:18:37,540
so you can subsidize
or you can mandate.

398
00:18:40,860 --> 00:18:41,610
Now these solutions--

399
00:18:41,610 --> 00:18:42,303
Yeah?

400
00:18:42,303 --> 00:18:46,247
AUDIENCE: Weren't a lot of
health insurance companies

401
00:18:46,247 --> 00:18:50,191
also looking at medical
records of

402
00:18:50,191 --> 00:18:52,163
people before deciding--

403
00:18:52,163 --> 00:18:55,121
were they unequally pricing
health care based

404
00:18:55,121 --> 00:18:57,093
on someone's history?

405
00:18:57,093 --> 00:19:00,051
And I know that caused a lot
of controversy say over

406
00:19:00,051 --> 00:19:01,037
pre-existing conditions.

407
00:19:01,037 --> 00:19:02,023
PROFESSOR: Yeah.

408
00:19:02,023 --> 00:19:03,400
Let me talk about
that next time.

409
00:19:03,400 --> 00:19:06,450
What you're asking is isn't it
in fact possible asymmetric

410
00:19:06,450 --> 00:19:08,170
information could flip?

411
00:19:08,170 --> 00:19:10,210
Isn't it possible insurers could
now know so much about

412
00:19:10,210 --> 00:19:12,030
us from public records that they
could start to have more

413
00:19:12,030 --> 00:19:14,490
information about us than we
even have about ourselves?

414
00:19:14,490 --> 00:19:16,800
And then they could decide to
get rid of us because they

415
00:19:16,800 --> 00:19:18,930
know we're a bad risk even
if we don't know.

416
00:19:18,930 --> 00:19:22,930
So could we be soon living in a
world of reverse asymmetric

417
00:19:22,930 --> 00:19:23,850
information?

418
00:19:23,850 --> 00:19:26,020
And that's possible, and I'll
talk about that next time.

419
00:19:26,020 --> 00:19:28,035
But for now let's assume we live
in a world where you know

420
00:19:28,035 --> 00:19:30,450
more than the insurer.

421
00:19:30,450 --> 00:19:32,220
So you can subsidize the guys,
you can mandate it.

422
00:19:32,220 --> 00:19:34,030
Now each of these have
pros and cons.

423
00:19:34,030 --> 00:19:38,310
If you subsidize them then the
MIT students are happy.

424
00:19:38,310 --> 00:19:40,880
They get insurance for free,
but MIT is out $10,000.

425
00:19:40,880 --> 00:19:42,070
Or if it's the government,
the government's

426
00:19:42,070 --> 00:19:43,320
out a lot of money.

427
00:19:43,320 --> 00:19:44,800
How does the government
raise that money?

428
00:19:44,800 --> 00:19:48,040
It taxes people, and that
has efficiency costs.

429
00:19:48,040 --> 00:19:48,970
So one thing we could
do is make health

430
00:19:48,970 --> 00:19:49,880
insurance free for everyone.

431
00:19:49,880 --> 00:19:51,900
That's what Canada does.

432
00:19:51,900 --> 00:19:54,610
Health insurance is
free for everyone.

433
00:19:54,610 --> 00:19:57,930
And the problem is to do that in
the US would cost something

434
00:19:57,930 --> 00:20:03,820
like a couple of trillion
dollars.

435
00:20:03,820 --> 00:20:07,530
That's a lot of money
to raise in taxes.

436
00:20:07,530 --> 00:20:10,090
And that would cause
inefficiency from raising that

437
00:20:10,090 --> 00:20:11,220
money in taxes.

438
00:20:11,220 --> 00:20:12,660
On the other hand the
mandate doesn't cost

439
00:20:12,660 --> 00:20:14,460
the government anything.

440
00:20:14,460 --> 00:20:18,560
But it makes healthy people very
upset because suddenly

441
00:20:18,560 --> 00:20:21,900
you're an MIT student, you know
for sure you're not going

442
00:20:21,900 --> 00:20:24,320
to spend zero but the government
mandates you have

443
00:20:24,320 --> 00:20:26,370
to pay that $100 premium.

444
00:20:26,370 --> 00:20:28,120
Well now you're upset.

445
00:20:28,120 --> 00:20:30,160
So basically there's no good
answer here to solve this

446
00:20:30,160 --> 00:20:31,920
asymmetric information
problem.

447
00:20:31,920 --> 00:20:34,120
Either we have to spend a lot
of money and then people

448
00:20:34,120 --> 00:20:35,600
aren't upset but
taxpayers are.

449
00:20:35,600 --> 00:20:36,950
Insurers aren't but
taxpayers are.

450
00:20:36,950 --> 00:20:38,690
Or we have to mandate people
to buy, in which case the

451
00:20:38,690 --> 00:20:41,270
healthy will be upset they're
being forced to buy something

452
00:20:41,270 --> 00:20:43,120
they don't want.

453
00:20:43,120 --> 00:20:45,200
And that's the difficult
position we face ourselves in

454
00:20:45,200 --> 00:20:46,980
with social insurance
programs.

455
00:20:46,980 --> 00:20:50,520
But the bottom line is what
we have is this failure of

456
00:20:50,520 --> 00:20:53,310
private insurance markets due to
asymmetric information and

457
00:20:53,310 --> 00:20:56,560
the need for the government to
come in and take one or both

458
00:20:56,560 --> 00:21:00,460
of these routes to
fix the problem.

459
00:21:00,460 --> 00:21:03,750
If the government doesn't then
you don't get people getting

460
00:21:03,750 --> 00:21:05,020
insurance that they'd
like to have. You

461
00:21:05,020 --> 00:21:07,246
get a market failure.

462
00:21:07,246 --> 00:21:09,850
You get a market failure in
people not having insurance

463
00:21:09,850 --> 00:21:11,980
that they would value.

464
00:21:11,980 --> 00:21:15,360
And that's why we have
governments come in and

465
00:21:15,360 --> 00:21:18,620
provide social insurance.

466
00:21:18,620 --> 00:21:20,815
Now, as I said, the government
comes in and provide many

467
00:21:20,815 --> 00:21:22,320
types of social insurance
of many types.

468
00:21:22,320 --> 00:21:24,870
There's medical social
insurance, and here the

469
00:21:24,870 --> 00:21:27,240
government does that through the
form of what we call two

470
00:21:27,240 --> 00:21:33,500
programs, Medicare
and Medicaid.

471
00:21:33,500 --> 00:21:38,700
Medicare is universal, heavily
subsidized health insurance

472
00:21:38,700 --> 00:21:41,230
for the elderly.

473
00:21:41,230 --> 00:21:44,800
Medicaid is like a
welfare program.

474
00:21:44,800 --> 00:21:47,710
It's means tested, categorical,
free health

475
00:21:47,710 --> 00:21:49,010
insurance for the poor.

476
00:21:49,010 --> 00:21:52,780
So Medicare's for the elderly,
Medicaid's for the poor.

477
00:21:52,780 --> 00:21:55,540
In neither case is there a
mandate, in both cases the

478
00:21:55,540 --> 00:21:57,420
government, essentially,
is making it free.

479
00:21:57,420 --> 00:22:00,530
Essentially it's gone the
subsidy route, as a result at

480
00:22:00,530 --> 00:22:03,370
an enormous cost to
the US taxpayer.

481
00:22:03,370 --> 00:22:07,740
Medicare is now about a $500
billion a year program.

482
00:22:07,740 --> 00:22:11,170
Medicaid is approaching a $400
billion a year program.

483
00:22:11,170 --> 00:22:14,060
It's almost a trillion dollars
US taxpayers are paying to get

484
00:22:14,060 --> 00:22:15,510
people to buy health
insurance.

485
00:22:15,510 --> 00:22:18,510
So by solving it with this route
these are incredibly,

486
00:22:18,510 --> 00:22:20,610
especially Medicare is an
incredibly politically popular

487
00:22:20,610 --> 00:22:22,080
program because why not?

488
00:22:22,080 --> 00:22:23,380
People are getting
it for free.

489
00:22:23,380 --> 00:22:26,280
But the cost is an enormous
budgetary pressure on the US

490
00:22:26,280 --> 00:22:27,620
government.

491
00:22:27,620 --> 00:22:30,760
That's for health.

492
00:22:30,760 --> 00:22:34,910
We also have disability
insurance, DI.

493
00:22:34,910 --> 00:22:37,880
That's for people who have
a career-ending injury.

494
00:22:37,880 --> 00:22:40,095
If you fall off your motorcycle
and get paralyzed

495
00:22:40,095 --> 00:22:43,360
and can't work anymore you can
get disability insurance,

496
00:22:43,360 --> 00:22:45,400
which is money which compensates
you for being

497
00:22:45,400 --> 00:22:47,035
unable to work having a
career-ending injury.

498
00:22:55,180 --> 00:22:57,750
So this is another form
of social insurance.

499
00:22:57,750 --> 00:23:02,430
We have workers' compensation,
which is insurance that you

500
00:23:02,430 --> 00:23:08,140
have to insure yourself against
injuries on the job.

501
00:23:08,140 --> 00:23:10,260
So if you slip and fall on the
job and have to miss a couple

502
00:23:10,260 --> 00:23:12,690
weeks of work this covers
your medical costs

503
00:23:12,690 --> 00:23:13,940
and your lost wages.

504
00:23:16,220 --> 00:23:18,720
And we have unemployment
insurance, which as I

505
00:23:18,720 --> 00:23:21,810
mentioned, is insurance for
people who lose their jobs due

506
00:23:21,810 --> 00:23:23,170
to economic circumstances.

507
00:23:23,170 --> 00:23:25,320
Unemployment insurance is not
for people who get fired

508
00:23:25,320 --> 00:23:26,295
because they're crappy
workers.

509
00:23:26,295 --> 00:23:28,570
It's for people who lose their
job because they're laid off

510
00:23:28,570 --> 00:23:29,990
for economic circumstances.

511
00:23:29,990 --> 00:23:34,700
They get some replacement of
wages under that unemployment

512
00:23:34,700 --> 00:23:35,450
insurance programs.

513
00:23:35,450 --> 00:23:37,260
These are the kind of social
insurance programs that we

514
00:23:37,260 --> 00:23:39,660
have in the US.

515
00:23:39,660 --> 00:23:42,960
And the arguments for them is
basically more complicated

516
00:23:42,960 --> 00:23:44,690
version of the arguments
I just made over there.

517
00:23:44,690 --> 00:23:46,780
The argument for all these
programs is that private

518
00:23:46,780 --> 00:23:49,680
insurance markets for things
like unemployment or injury

519
00:23:49,680 --> 00:23:52,590
will not function well due to
asymmetric information so the

520
00:23:52,590 --> 00:23:56,890
government needs to come in and
provide that insurance.

521
00:23:56,890 --> 00:24:00,220
But at the same time all of
these insurance programs have

522
00:24:00,220 --> 00:24:06,320
a cost. Not just the cost of
raising the money but the cost

523
00:24:06,320 --> 00:24:14,000
of what we call moral hazard,
which is that when you insure

524
00:24:14,000 --> 00:24:17,200
people against adverse risk you

525
00:24:17,200 --> 00:24:18,805
encourage adverse behavior.

526
00:24:21,910 --> 00:24:24,890
When you insure people against
adverse risk you encourage

527
00:24:24,890 --> 00:24:27,460
adverse behavior.

528
00:24:27,460 --> 00:24:30,410
And the problem here is another
form of asymmetric

529
00:24:30,410 --> 00:24:31,800
information.

530
00:24:31,800 --> 00:24:35,640
So one problem of asymmetric
information is since insurers

531
00:24:35,640 --> 00:24:37,790
don't know enough about you
they'll under insure you.

532
00:24:37,790 --> 00:24:39,540
So it's a market failure.

533
00:24:39,540 --> 00:24:42,500
This is a flip side market
failure, which is since

534
00:24:42,500 --> 00:24:45,790
insurers can't perfectly be sure
about what's wrong with

535
00:24:45,790 --> 00:24:50,510
you, you will be encouraged to
misrepresent your state in

536
00:24:50,510 --> 00:24:53,540
order to collect money.

537
00:24:53,540 --> 00:24:56,270
So a classic example
here is workers'

538
00:24:56,270 --> 00:24:58,950
compensation insurance.

539
00:24:58,950 --> 00:25:01,000
Workers' compensation insurance
is about a $50

540
00:25:01,000 --> 00:25:03,866
billion a year program that
provides money to people who

541
00:25:03,866 --> 00:25:06,140
are injured on the job.

542
00:25:06,140 --> 00:25:10,850
The problem is injured on the
job is very hard to verify.

543
00:25:10,850 --> 00:25:13,770
Most of the injuries are
not losing an arm.

544
00:25:13,770 --> 00:25:16,090
Most of the injuries are
straining your back.

545
00:25:16,090 --> 00:25:18,350
As anyone with back problems
knows it's pretty darn hard to

546
00:25:18,350 --> 00:25:21,230
verify if your back's
hurt or not.

547
00:25:21,230 --> 00:25:23,940
My first job in my first summer
after MIT was I was a

548
00:25:23,940 --> 00:25:26,820
filing clerk at a
doctor's office.

549
00:25:26,820 --> 00:25:28,830
The doctor did workers' comp--
this is a real story.

550
00:25:28,830 --> 00:25:32,630
Doctors do workers' comp claims.
It was one doctor with

551
00:25:32,630 --> 00:25:35,820
a staff of five secretaries and
an enormous filing staff.

552
00:25:35,820 --> 00:25:38,540
And the reason was the doctor
saw a patient every three

553
00:25:38,540 --> 00:25:40,270
minutes all day.

554
00:25:40,270 --> 00:25:42,060
They'd come in, he'd say, yeah,
you're hurt, here's your

555
00:25:42,060 --> 00:25:45,310
chit for workers' comp,
off you go.

556
00:25:45,310 --> 00:25:47,880
All day.

557
00:25:47,880 --> 00:25:50,460
So basically, because it's
incredibly hard to verify,

558
00:25:50,460 --> 00:25:52,260
this doctor was just-- basically
his role was saying

559
00:25:52,260 --> 00:25:55,700
yes to people so they could
get out of work.

560
00:25:55,700 --> 00:25:58,850
This is the problem we
have with a system--

561
00:25:58,850 --> 00:26:03,400
this is another information
asymmetry, which is now you,

562
00:26:03,400 --> 00:26:06,470
the insured, know more about
what's wrong with you than the

563
00:26:06,470 --> 00:26:08,670
insurer does.

564
00:26:08,670 --> 00:26:10,760
But it's a different form of
information asymmetry.

565
00:26:10,760 --> 00:26:14,670
Now it's that you may actually
not be as injured as you claim

566
00:26:14,670 --> 00:26:15,720
or need insurance
as you claim.

567
00:26:15,720 --> 00:26:17,920
As a result you over claim
insurance, and that costs the

568
00:26:17,920 --> 00:26:20,710
insurance industry money.

569
00:26:20,710 --> 00:26:24,880
Now, moral hazard,
theoretically, is unambiguous.

570
00:26:24,880 --> 00:26:27,310
But obviously it's going to
vary across situations.

571
00:26:27,310 --> 00:26:29,440
There's not going to be a whole
lot of moral hazard in

572
00:26:29,440 --> 00:26:31,170
programs that insure you
against blindness.

573
00:26:35,570 --> 00:26:37,770
And in that case that's because
it's pretty easy to

574
00:26:37,770 --> 00:26:39,230
observe if someone's
really blind.

575
00:26:39,230 --> 00:26:41,530
There's not a whole lot of
information asymmetry.

576
00:26:41,530 --> 00:26:43,520
There's going to be a lot of
moral hazard in programs that

577
00:26:43,520 --> 00:26:47,230
insure you against having a bad
back because that's harder

578
00:26:47,230 --> 00:26:48,790
to observe.

579
00:26:48,790 --> 00:26:52,250
And in fact we have lots of
evidence that moral hazard's a

580
00:26:52,250 --> 00:26:55,640
big problem with social
insurance programs. So one

581
00:26:55,640 --> 00:26:58,560
type of evidence is sort
of fun stories.

582
00:26:58,560 --> 00:27:01,010
So there's stories about, for
example, the Massachusetts

583
00:27:01,010 --> 00:27:04,870
prison guard who got about
$25,000 in workers' comp

584
00:27:04,870 --> 00:27:07,890
claims because he claimed he
hurt his back at work until

585
00:27:07,890 --> 00:27:09,650
people went online in order to
see that the same guy was

586
00:27:09,650 --> 00:27:10,570
running a karate school.

587
00:27:10,570 --> 00:27:13,130
And there was videos of him
doing karate moves and stuff

588
00:27:13,130 --> 00:27:16,000
online running his
karate school.

589
00:27:16,000 --> 00:27:17,690
Or the other guy, who
unfortunately had good

590
00:27:17,690 --> 00:27:20,980
instincts, he was on workers'
comp because he was injured.

591
00:27:20,980 --> 00:27:23,250
But then there was footage of
him at 9/11 flying down and

592
00:27:23,250 --> 00:27:24,970
carrying buckets back and forth
and helping put out the

593
00:27:24,970 --> 00:27:27,110
fire, which was a nice thing
for him to do, but clearly

594
00:27:27,110 --> 00:27:28,940
illustrated that he was not
too injured to be able to

595
00:27:28,940 --> 00:27:30,500
carry out his job.

596
00:27:30,500 --> 00:27:33,880
So basically, there's lots of
observational evidence.

597
00:27:33,880 --> 00:27:38,870
But there's also very clear
statistical evidence, which is

598
00:27:38,870 --> 00:27:43,420
if moral hazard doesn't exist
then when the program gets

599
00:27:43,420 --> 00:27:46,420
more generous I shouldn't
be more injured, right?

600
00:27:46,420 --> 00:27:47,430
My injury should be
something that

601
00:27:47,430 --> 00:27:49,160
happens or doesn't happen.

602
00:27:49,160 --> 00:27:52,480
But if more generous programs
lead to more injuries, or

603
00:27:52,480 --> 00:27:55,580
people staying out of work
longer, or people using more

604
00:27:55,580 --> 00:28:00,120
health care then that suggests
moral hazard.

605
00:28:00,120 --> 00:28:02,460
The generosity, whether you're
sick or not, shouldn't be

606
00:28:02,460 --> 00:28:04,610
affected by the generosity
of health insurance.

607
00:28:04,610 --> 00:28:06,320
But if you suddenly see more
generous health insurance

608
00:28:06,320 --> 00:28:07,890
leads to more people
being sick that

609
00:28:07,890 --> 00:28:09,240
suggests moral hazard.

610
00:28:09,240 --> 00:28:12,510
And there's hundreds of studies
of this type showing,

611
00:28:12,510 --> 00:28:15,880
in all these realms, enormous
moral hazard.

612
00:28:15,880 --> 00:28:18,950
Showing, for example, for
workers' comp, that when you

613
00:28:18,950 --> 00:28:21,890
make workers' comp more generous
people report more

614
00:28:21,890 --> 00:28:24,570
injuries on the job.

615
00:28:24,570 --> 00:28:27,160
Here's my favorite
fact about that.

616
00:28:27,160 --> 00:28:31,980
Someone did a study of type of
injury reported to workers'

617
00:28:31,980 --> 00:28:35,170
comp by day of the week.

618
00:28:35,170 --> 00:28:39,570
And what they found was on
Mondays compared to other days

619
00:28:39,570 --> 00:28:42,730
of the week there was a very
high share of sprains and

620
00:28:42,730 --> 00:28:47,290
strains and a low share of
lacerations, of cuts.

621
00:28:47,290 --> 00:28:51,340
So every other day of the week,
say it was 50% cuts and

622
00:28:51,340 --> 00:28:52,570
50% strains.

623
00:28:52,570 --> 00:28:56,140
On Monday it was 75% strains
and 25% cuts.

624
00:28:56,140 --> 00:28:57,365
Why is that?

625
00:28:57,365 --> 00:28:59,590
Well, that's because guys got
hurt over the weekend playing

626
00:28:59,590 --> 00:29:02,390
softball and they came in on
Monday and they said they got

627
00:29:02,390 --> 00:29:03,000
hurt on the job.

628
00:29:03,000 --> 00:29:04,700
Now you can't do that if you cut
yourself on the weekend in

629
00:29:04,700 --> 00:29:07,630
your shop because it's scabbed
over by Monday.

630
00:29:07,630 --> 00:29:09,450
But if you hurt your back on
the weekend you can drag

631
00:29:09,450 --> 00:29:11,920
yourself over and say, ooh,
I hurt my back at work.

632
00:29:11,920 --> 00:29:13,610
So that's exactly the kind of
evidence that's consistent

633
00:29:13,610 --> 00:29:14,900
with moral hazard.

634
00:29:14,900 --> 00:29:16,500
And we see that everywhere.

635
00:29:16,500 --> 00:29:18,170
There's huge amounts
of moral hazard.

636
00:29:18,170 --> 00:29:20,230
Moral hazard, for example, in
health insurance the best

637
00:29:20,230 --> 00:29:23,920
estimates are that the amount
of extra money that we spend

638
00:29:23,920 --> 00:29:26,500
on society on health care
because of moral hazard on the

639
00:29:26,500 --> 00:29:29,580
order of $500 billion a year in
terms of extra care that's

640
00:29:29,580 --> 00:29:31,360
used because people
are insured.

641
00:29:31,360 --> 00:29:33,160
That does nothing to improve
their health, they just use it

642
00:29:33,160 --> 00:29:34,020
because they're insured.

643
00:29:34,020 --> 00:29:37,540
It's on the order of $500
billion a year.

644
00:29:37,540 --> 00:29:39,440
This is an enormous problem.

645
00:29:39,440 --> 00:29:42,700
So here we have what we call
the classic social

646
00:29:42,700 --> 00:29:44,760
insurance trade off.

647
00:29:44,760 --> 00:29:47,840
On the one hand, people
value insurance.

648
00:29:47,840 --> 00:29:52,880
Insurance is very valuable to
help you in the state when you

649
00:29:52,880 --> 00:29:58,270
get in an accident or are
injured in some way.

650
00:29:58,270 --> 00:30:01,360
On the other hand, giving people
insurance against those

651
00:30:01,360 --> 00:30:06,010
bad outcomes will encourage
bad behavior.

652
00:30:06,010 --> 00:30:10,070
And that bad behavior has
two forms of costs.

653
00:30:10,070 --> 00:30:13,080
There's two costs to
that bad behavior.

654
00:30:13,080 --> 00:30:18,700
The first cost is the tax costs
and the deadweight loss

655
00:30:18,700 --> 00:30:22,270
of taxation, which is if there's
$500 billion extra in

656
00:30:22,270 --> 00:30:27,040
medical care used because people
are insured then that

657
00:30:27,040 --> 00:30:31,170
means that's $500 billion of
taxes we have to raise, or

658
00:30:31,170 --> 00:30:33,193
$500 billion in resource
costs to society.

659
00:30:33,193 --> 00:30:34,230
Well, not all that's
government.

660
00:30:34,230 --> 00:30:36,420
But a couple hundred billion
in taxes we have to raise,

661
00:30:36,420 --> 00:30:38,610
that causes deadweight loss.

662
00:30:38,610 --> 00:30:42,070
The more important cost is
really the overall efficiency

663
00:30:42,070 --> 00:30:48,110
loss from people using resources
inefficiently, from

664
00:30:48,110 --> 00:30:51,430
people not behaving optimally
because of the incentives of

665
00:30:51,430 --> 00:30:54,270
social insurance.

666
00:30:54,270 --> 00:30:56,290
So let's go back and think about
the case of workers'

667
00:30:56,290 --> 00:31:00,780
comp, and let's think about
our standard leisure

668
00:31:00,780 --> 00:31:03,320
consumption trade off.

669
00:31:03,320 --> 00:31:07,090
We have our standard leisure
consumption trade off where

670
00:31:07,090 --> 00:31:10,370
you have leisure on this axis,
consumption on this axis, some

671
00:31:10,370 --> 00:31:12,773
budget constraint, and people
choose how much leisure to

672
00:31:12,773 --> 00:31:15,540
take and how much consumption
to take.

673
00:31:15,540 --> 00:31:21,460
Now the slope of this budget
constraint is minus your wage.

674
00:31:21,460 --> 00:31:24,110
That's the price of leisure.

675
00:31:24,110 --> 00:31:27,670
So the price of sitting at home
is the wage you could

676
00:31:27,670 --> 00:31:30,390
have earned out in the market.

677
00:31:30,390 --> 00:31:36,220
What this means is you will sit
at home until the value

678
00:31:36,220 --> 00:31:39,630
you get from sitting at home
falls below the value you get

679
00:31:39,630 --> 00:31:40,710
out in the market.

680
00:31:40,710 --> 00:31:43,220
So in other words, you wouldn't
work 24 hours a day

681
00:31:43,220 --> 00:31:46,680
because that 24th hour, the
sleep would be so valuable

682
00:31:46,680 --> 00:31:50,920
that it would clearly exceed
the wage you could earn.

683
00:31:50,920 --> 00:31:52,820
So when you decide how hard to
work, when an individual

684
00:31:52,820 --> 00:31:53,310
decides how--

685
00:31:53,310 --> 00:31:56,090
presuming we can freely
choose our hours--

686
00:31:56,090 --> 00:32:01,730
we decide how hard to work, we
basically trade off how much

687
00:32:01,730 --> 00:32:05,170
we value not working versus
how much we'd

688
00:32:05,170 --> 00:32:05,970
earn if we do work.

689
00:32:05,970 --> 00:32:08,670
If you're in a high wage you
might work more hours, if

690
00:32:08,670 --> 00:32:10,583
you're in a low wage you'll earn
less because if you're in

691
00:32:10,583 --> 00:32:14,320
a low wage you might as well sit
at home and watch Oprah.

692
00:32:14,320 --> 00:32:15,920
If you earn a high wage you're
not going to watch Oprah

693
00:32:15,920 --> 00:32:17,560
you're going to go
earn the money.

694
00:32:17,560 --> 00:32:19,410
And that's how we make our
labor supply decisions.

695
00:32:19,410 --> 00:32:25,340
We set the wage equal to the
marginal value of leisure.

696
00:32:25,340 --> 00:32:29,650
That's how we make our labor
decisions, is we say if the

697
00:32:29,650 --> 00:32:32,190
marginal value of the next hour
of leisure exceeds my

698
00:32:32,190 --> 00:32:35,400
wage I'll stay home.

699
00:32:35,400 --> 00:32:37,970
If the marginal value of the
next hour of leisure is below

700
00:32:37,970 --> 00:32:39,220
my wage I'll going to work.

701
00:32:41,530 --> 00:32:43,620
And that's how we make our
labor supply decisions.

702
00:32:43,620 --> 00:32:46,900
What does workers' comp do
to this, or UI or other

703
00:32:46,900 --> 00:32:48,310
programs like that?

704
00:32:48,310 --> 00:32:51,690
What that says is wait a second,
now if I sit at home I

705
00:32:51,690 --> 00:32:54,760
don't just get the marginal
value of my leisure, I also

706
00:32:54,760 --> 00:32:56,780
get a check from
the government.

707
00:32:56,780 --> 00:32:59,830
So now if I sit at home I get
the marginal value of my

708
00:32:59,830 --> 00:33:02,790
leisure plus a workers'
comp check.

709
00:33:02,790 --> 00:33:05,930
Whereas if I work I still
just get my wage.

710
00:33:05,930 --> 00:33:07,470
So what's that going to do to
the amount of hours I'm going

711
00:33:07,470 --> 00:33:10,110
to want to work?

712
00:33:10,110 --> 00:33:13,610
It's going to reduce them
because now it's makes sitting

713
00:33:13,610 --> 00:33:16,570
at home more attractive because
now I'm going to work

714
00:33:16,570 --> 00:33:19,370
only until my wage exceeds my
marginal value of leisure plus

715
00:33:19,370 --> 00:33:21,580
workers' comp.

716
00:33:21,580 --> 00:33:23,570
So if workers' comp is big
enough and my wage isn't that

717
00:33:23,570 --> 00:33:27,530
high, in the limit, if workers
comp replaced all my wages I

718
00:33:27,530 --> 00:33:29,180
would never work.

719
00:33:29,180 --> 00:33:38,710
If workers' comp had what we
call a 100% replacement rate,

720
00:33:38,710 --> 00:33:42,260
replacement rate, if it replaced
100% of my wages I

721
00:33:42,260 --> 00:33:43,040
would never work.

722
00:33:43,040 --> 00:33:43,430
Why?

723
00:33:43,430 --> 00:33:45,760
Because by definition this side
would always be less than

724
00:33:45,760 --> 00:33:50,300
this side, so I'd never work.

725
00:33:50,300 --> 00:33:52,620
And that's the efficiency loss
of these social insurance

726
00:33:52,620 --> 00:33:55,550
programs. That people are not
engaging in productive trades.

727
00:33:55,550 --> 00:33:57,930
Society just cares
about this part.

728
00:33:57,930 --> 00:34:00,700
Society just cares about your
wage versus your marginal

729
00:34:00,700 --> 00:34:01,190
value of leisure.

730
00:34:01,190 --> 00:34:03,790
So if you work less than that
because of this part then

731
00:34:03,790 --> 00:34:06,550
that's a distortion, that's an
efficiency cost to society.

732
00:34:06,550 --> 00:34:08,810
You are sitting at home when
you should be making a

733
00:34:08,810 --> 00:34:13,280
productive trade of your
labor for output goods.

734
00:34:13,280 --> 00:34:15,320
And that's the efficiency cost
of social insurance.

735
00:34:17,989 --> 00:34:19,739
So what do we do?

736
00:34:19,739 --> 00:34:22,420
We want social insurance, on the
one hand, because private

737
00:34:22,420 --> 00:34:25,070
insurance markets fail.

738
00:34:25,070 --> 00:34:27,380
We don't want social insurance,
on the other hand,

739
00:34:27,380 --> 00:34:29,440
because people are going
to sit at home.

740
00:34:29,440 --> 00:34:30,110
What do we do?

741
00:34:30,110 --> 00:34:32,800
Well the answer is
we compromise.

742
00:34:32,800 --> 00:34:36,489
We have social insurance but
don't make it that generous.

743
00:34:36,489 --> 00:34:38,980
So for example, the unemployment
insurance

744
00:34:38,980 --> 00:34:44,060
program, that's insurance that
is valuable because there's no

745
00:34:44,060 --> 00:34:45,840
way to get private insurance
against losing your job.

746
00:34:45,840 --> 00:34:47,420
I'd never want to sell
that product.

747
00:34:47,420 --> 00:34:49,380
We know who'd take
that, right?

748
00:34:49,380 --> 00:34:50,989
On the other hand, we know
there's moral hazard.

749
00:34:50,989 --> 00:34:53,610
We know that people sit at home
longer rather looking for

750
00:34:53,610 --> 00:34:55,830
a job because they
get UI checks.

751
00:34:55,830 --> 00:34:56,850
There's great evidence
of that.

752
00:34:56,850 --> 00:34:59,710
For example, if you look at how
long people sit at home,

753
00:34:59,710 --> 00:35:01,920
everyone sits at home and then
suddenly gets a job the week

754
00:35:01,920 --> 00:35:03,710
UI runs out.

755
00:35:03,710 --> 00:35:06,020
That's an exaggeration, but
roughly speaking, there's a

756
00:35:06,020 --> 00:35:07,680
huge amount of people
sitting at home

757
00:35:07,680 --> 00:35:09,090
collecting the UI checks.

758
00:35:09,090 --> 00:35:10,150
What do we do?

759
00:35:10,150 --> 00:35:13,660
Well the UI replacement
rate is around 50%.

760
00:35:13,660 --> 00:35:15,230
We say, look, we know you're
going to want to sit home so

761
00:35:15,230 --> 00:35:16,540
we're going to make your
life uncomfortable.

762
00:35:16,540 --> 00:35:20,310
We'll only give you half of what
you'd earn at work, but

763
00:35:20,310 --> 00:35:21,780
we don't want you to starve so
we'll give you half of what

764
00:35:21,780 --> 00:35:23,280
you earn at work.

765
00:35:23,280 --> 00:35:26,290
And so the replacement rate
is chosen to trade off the

766
00:35:26,290 --> 00:35:29,800
benefits of insuring people
against these bad events

767
00:35:29,800 --> 00:35:31,880
against the costs of distorting
their behavior

768
00:35:31,880 --> 00:35:33,010
through moral hazard.

769
00:35:33,010 --> 00:35:34,360
And that's the social
insurance trade off.

770
00:35:34,360 --> 00:35:37,150
That's how we set up these
social insurance programs.

771
00:35:37,150 --> 00:35:40,710
They sort of embody
that trade off.

772
00:35:40,710 --> 00:35:41,960
Questions about that?

773
00:35:44,090 --> 00:35:47,100
Now, so what I want to do in the
remaining time is I want

774
00:35:47,100 --> 00:35:49,830
to talk about the nation's
largest single social

775
00:35:49,830 --> 00:35:53,130
insurance program, one you've
heard about a lot, which is

776
00:35:53,130 --> 00:35:53,685
Social Security.

777
00:35:53,685 --> 00:35:55,840
We're going to talk about
Social Security.

778
00:35:55,840 --> 00:35:59,650
I'm going to come back next
lecture and we'll go over a

779
00:35:59,650 --> 00:36:01,850
bit what's going to be on
the final in the last 15

780
00:36:01,850 --> 00:36:02,520
minutes next time.

781
00:36:02,520 --> 00:36:04,550
But the first part of the
lecture I want to talk about

782
00:36:04,550 --> 00:36:07,030
health insurance, which is our
single largest government

783
00:36:07,030 --> 00:36:08,320
expenditure.

784
00:36:08,320 --> 00:36:10,980
But the single biggest
government program right now,

785
00:36:10,980 --> 00:36:13,890
social insurance program, it
remains Social Security.

786
00:36:13,890 --> 00:36:16,600
Medicare is going to pass it
in about two years, but for

787
00:36:16,600 --> 00:36:17,890
now it's still Social
Security.

788
00:36:20,750 --> 00:36:22,025
What is Social Security?

789
00:36:22,025 --> 00:36:27,430
Social Security is insurance
against the income loss from

790
00:36:27,430 --> 00:36:30,220
retirement.

791
00:36:30,220 --> 00:36:33,930
We know that when you have to
retire you're going to suffer

792
00:36:33,930 --> 00:36:36,180
a big income loss.

793
00:36:36,180 --> 00:36:39,000
What the government does is
insures you against that

794
00:36:39,000 --> 00:36:42,270
income loss by providing
you Social Security.

795
00:36:42,270 --> 00:36:44,170
And here's how it works.

796
00:36:44,170 --> 00:36:46,870
How many of you have seen a pay
stub and you've seen the

797
00:36:46,870 --> 00:36:48,950
word FICA on it, and then
there's a charge that's coming

798
00:36:48,950 --> 00:36:50,320
out of your pay?

799
00:36:50,320 --> 00:36:53,180
FICA is the Federal Insurance
Contribution Amount.

800
00:36:53,180 --> 00:36:56,610
That's the money you're paying
in to support the social

801
00:36:56,610 --> 00:36:59,210
insurance program.

802
00:36:59,210 --> 00:37:02,210
You pay that in, and that's
currently 12.4% of wages, well

803
00:37:02,210 --> 00:37:05,325
6.2% on you and 6.2%
on your employer.

804
00:37:05,325 --> 00:37:08,950
You each pay 6.2%
of your wages.

805
00:37:08,950 --> 00:37:12,820
That goes into a trust fund and
that money then gets paid

806
00:37:12,820 --> 00:37:15,000
out to retirees, and hopefully
paid out to

807
00:37:15,000 --> 00:37:16,250
you when you retire.

808
00:37:18,130 --> 00:37:22,160
So the way Social Security works
is you pay in money when

809
00:37:22,160 --> 00:37:25,470
you're young, that money
gets paid out to

810
00:37:25,470 --> 00:37:28,470
you when you retire.

811
00:37:28,470 --> 00:37:30,550
Not that same money because
your money going in now is

812
00:37:30,550 --> 00:37:31,820
going to today's retirees.

813
00:37:31,820 --> 00:37:34,820
Future workers will pay in and
that money will go to you.

814
00:37:34,820 --> 00:37:35,970
And the Social Security
replacement

815
00:37:35,970 --> 00:37:38,630
rate is around 50%.

816
00:37:38,630 --> 00:37:40,280
So to try to deal with this
problem the Social Security

817
00:37:40,280 --> 00:37:46,880
replacement rate's around 50%.

818
00:37:46,880 --> 00:37:48,320
So what's the benefit
of Social Security?

819
00:37:48,320 --> 00:37:50,730
Well, clearly the benefit is
people don't want to starve

820
00:37:50,730 --> 00:37:52,130
when they retire.

821
00:37:52,130 --> 00:37:53,610
We'd like insurance
against the income

822
00:37:53,610 --> 00:37:56,200
loss of when we retire.

823
00:37:56,200 --> 00:37:58,940
Now, our own savings can do that
to some extent but it's

824
00:37:58,940 --> 00:38:01,160
not perfect because you don't
know exactly how much to save,

825
00:38:01,160 --> 00:38:02,340
you don't know when you retire,
you don't know how

826
00:38:02,340 --> 00:38:03,580
long you'll live.

827
00:38:03,580 --> 00:38:04,800
So the government
social insurance

828
00:38:04,800 --> 00:38:06,500
program helps with that.

829
00:38:06,500 --> 00:38:10,320
But there's also a moral hazard
cost, which is if you

830
00:38:10,320 --> 00:38:13,180
insure people against retirement
you encourage

831
00:38:13,180 --> 00:38:17,970
retirement, because now we're
saying is if you work you get

832
00:38:17,970 --> 00:38:21,530
your wage but if you retire you
get your Social Security.

833
00:38:21,530 --> 00:38:26,750
So now, you're eligible for
Social Security at age 62.

834
00:38:26,750 --> 00:38:30,310
So at age 62 you're suddenly
eligible for about half your

835
00:38:30,310 --> 00:38:31,980
wage in Social Security.

836
00:38:31,980 --> 00:38:35,250
So now your decision
is, gee, do I work?

837
00:38:35,250 --> 00:38:37,250
Well that's contrasted against
my marginal value of leisure

838
00:38:37,250 --> 00:38:39,690
plus half my wage.

839
00:38:39,690 --> 00:38:41,730
So unless the marginal value of
leisure is less than half

840
00:38:41,730 --> 00:38:45,040
my wage I'm going to quit.

841
00:38:45,040 --> 00:38:49,230
So what Social Security is
doing is distorting your

842
00:38:49,230 --> 00:38:51,450
decision and causing you
to retire early.

843
00:38:51,450 --> 00:38:54,950
That's the moral hazard part.

844
00:38:54,950 --> 00:38:58,500
Now in the US we recognize
this, and we do

845
00:38:58,500 --> 00:39:00,342
sort of a neat trick.

846
00:39:00,342 --> 00:39:04,640
What we do in US is we have
something called an actuarial

847
00:39:04,640 --> 00:39:10,210
adjustment to your Social
Security benefits.

848
00:39:10,210 --> 00:39:12,930
What that means is the
following, you're 62 and you

849
00:39:12,930 --> 00:39:17,160
say, wait a second, I can retire
now, and if I retire

850
00:39:17,160 --> 00:39:19,920
now I'm going to get
half my wage.

851
00:39:19,920 --> 00:39:20,920
So why not retire now?

852
00:39:20,920 --> 00:39:22,910
We say, wait a second, we've
got a deal for you.

853
00:39:22,910 --> 00:39:27,130
If you work one more year and
retire next year then we'll

854
00:39:27,130 --> 00:39:30,170
give you 7% more every year
for the rest of your life.

855
00:39:30,170 --> 00:39:33,040
Your check will be 7% higher
for the rest your life.

856
00:39:33,040 --> 00:39:35,473
If you work two more years it
will be 14% higher, 3 more

857
00:39:35,473 --> 00:39:38,180
years 21% higher, et cetera.

858
00:39:38,180 --> 00:39:41,290
We recognize the moral hazard,
and we raise the check the

859
00:39:41,290 --> 00:39:45,260
longer you work to
make up for it.

860
00:39:45,260 --> 00:39:47,960
So the US recognizes that, and
as a result there's not a huge

861
00:39:47,960 --> 00:39:50,310
moral hazard from Social
Security in the US.

862
00:39:50,310 --> 00:39:57,590
It's pretty modest. In Europe
this is not the case.

863
00:39:57,590 --> 00:39:58,800
Let's take the country--

864
00:39:58,800 --> 00:39:59,800
let me back up.

865
00:39:59,800 --> 00:40:01,980
So what the actuarial adjustment
does is says we

866
00:40:01,980 --> 00:40:03,220
recognize that Social
Security is

867
00:40:03,220 --> 00:40:05,310
essentially taxing your work.

868
00:40:05,310 --> 00:40:08,250
To offset that we're going to
subsidize you to work more by

869
00:40:08,250 --> 00:40:12,520
raising your Social Security
check the more you work.

870
00:40:12,520 --> 00:40:15,940
In Europe they're not
as smart as us.

871
00:40:15,940 --> 00:40:18,170
They don't recognize this, or
if they do they're not smart

872
00:40:18,170 --> 00:40:20,120
enough to reflect it.

873
00:40:20,120 --> 00:40:21,480
They don't have an actuarial
adjustment.

874
00:40:21,480 --> 00:40:23,200
So here's how the Social
Security system works in the

875
00:40:23,200 --> 00:40:24,690
Netherlands, for example.

876
00:40:24,690 --> 00:40:28,550
When you reach 55 in the
Netherlands your choice is you

877
00:40:28,550 --> 00:40:31,660
can work or you can stop working
and get a Social

878
00:40:31,660 --> 00:40:35,480
Security check which is
90% of your wage.

879
00:40:35,480 --> 00:40:38,210
So literally, your choice is
work or sit at home and get

880
00:40:38,210 --> 00:40:41,960
90% of your wage with no
actuarial adjustment.

881
00:40:41,960 --> 00:40:44,530
So unless the value of your
leisure is less than 10% of

882
00:40:44,530 --> 00:40:46,640
your wage you can just
sit at home.

883
00:40:46,640 --> 00:40:49,370
But wait, that's not all.

884
00:40:49,370 --> 00:40:51,510
This is a very generous
program, right?

885
00:40:51,510 --> 00:40:52,880
How do they pay for that?

886
00:40:52,880 --> 00:40:55,500
They pay for it by
taxing workers.

887
00:40:55,500 --> 00:40:59,290
So if you work you get your wage
but then you pay a 40%

888
00:40:59,290 --> 00:41:01,280
payroll tax.

889
00:41:01,280 --> 00:41:05,320
So actually, you can work and
get 60% of your wage or sit at

890
00:41:05,320 --> 00:41:07,970
home and get 90% of your wage.

891
00:41:07,970 --> 00:41:08,640
Guess what?

892
00:41:08,640 --> 00:41:13,140
No one works after age 55
in the Netherlands.

893
00:41:13,140 --> 00:41:15,660
Literally.

894
00:41:15,660 --> 00:41:17,060
Now people will do
some underground

895
00:41:17,060 --> 00:41:17,970
work that's not taxed.

896
00:41:17,970 --> 00:41:20,360
They'll paint houses or things
which they can do on the

897
00:41:20,360 --> 00:41:21,390
underground economy.

898
00:41:21,390 --> 00:41:25,030
But literally, nobody works
after 55 in the Netherlands.

899
00:41:25,030 --> 00:41:26,035
Economics works.

900
00:41:26,035 --> 00:41:30,880
When you make the incentives
that blatant economics works.

901
00:41:30,880 --> 00:41:32,120
And that's the moral hazard.

902
00:41:32,120 --> 00:41:35,830
That's an example of not
respecting the social

903
00:41:35,830 --> 00:41:39,220
insurance trade off, erring on
the side of insuring people

904
00:41:39,220 --> 00:41:42,340
over respecting the problems
of moral hazard.

905
00:41:42,340 --> 00:41:45,720
And the answer is part of the
major problem that they have

906
00:41:45,720 --> 00:41:49,120
in these countries, part of the
major deficit, prolonged

907
00:41:49,120 --> 00:41:52,730
deficit problem they face is
that older people don't work.

908
00:41:52,730 --> 00:41:56,530
Now these countries have started
to recognize this, and

909
00:41:56,530 --> 00:41:57,580
they're starting to try
to deal with this.

910
00:41:57,580 --> 00:41:59,950
In fact, the country of France,
you may have noticed,

911
00:41:59,950 --> 00:42:01,530
recently tried to
deal with this.

912
00:42:01,530 --> 00:42:04,520
They have proposed raising their
retirement age, which is

913
00:42:04,520 --> 00:42:07,170
age 60, beyond which,
essentially, no one works in

914
00:42:07,170 --> 00:42:11,180
France, to age 62.

915
00:42:11,180 --> 00:42:11,710
What happened?

916
00:42:11,710 --> 00:42:15,470
You may have heard, massive
strikes, enormous political

917
00:42:15,470 --> 00:42:19,160
strife because people don't
understand economics.

918
00:42:19,160 --> 00:42:22,912
People don't understand that,
gee, yes that'll be a bummer

919
00:42:22,912 --> 00:42:25,160
if you work a couple years
longer, but that's what's

920
00:42:25,160 --> 00:42:26,930
going to make our
economy fiscally

921
00:42:26,930 --> 00:42:28,720
sound and more efficient.

922
00:42:28,720 --> 00:42:31,550
And in the long run we all
benefit from that.

923
00:42:31,550 --> 00:42:33,410
So the government's, by the
way, not proposing that

924
00:42:33,410 --> 00:42:36,300
today's retirees are hurt,
saying starting in 20 years

925
00:42:36,300 --> 00:42:37,815
you have to work two
years longer.

926
00:42:37,815 --> 00:42:41,100
So overall everyone's going to
benefit, but the bottom line

927
00:42:41,100 --> 00:42:42,570
is people are too short-sighted
to understand

928
00:42:42,570 --> 00:42:45,130
the economics, and that leads
to the rioting and things we

929
00:42:45,130 --> 00:42:47,310
saw in France.

930
00:42:47,310 --> 00:42:50,240
This stuff is hard to
a normal person.

931
00:42:50,240 --> 00:42:53,405
You guys are smart and you spent
a semester learning it.

932
00:42:53,405 --> 00:42:55,360
If you take a regular person in
the street they just don't

933
00:42:55,360 --> 00:42:58,890
understand this trade off,
and that's the problem.

934
00:42:58,890 --> 00:43:00,980
That's the problem in trying
to deal with these moral

935
00:43:00,980 --> 00:43:05,120
hazard issues by making social
insurance less generous.

936
00:43:05,120 --> 00:43:07,420
Now, let me talk about one
other issue with Social

937
00:43:07,420 --> 00:43:09,310
Security because it's been in
the news and will be in the

938
00:43:09,310 --> 00:43:12,760
news, which is the fact that
Social Security is

939
00:43:12,760 --> 00:43:14,870
running out of money.

940
00:43:14,870 --> 00:43:17,730
Now, Social Security is running
out of money because

941
00:43:17,730 --> 00:43:19,090
here's a dirty secret
about Social

942
00:43:19,090 --> 00:43:20,840
Security you may not know.

943
00:43:20,840 --> 00:43:23,550
How many of you have ever
heard of a Ponzi scheme?

944
00:43:23,550 --> 00:43:24,515
Anyone watch Boardwalk Empire?

945
00:43:24,515 --> 00:43:26,810
They're doing that now,
Boardwalk Empire.

946
00:43:26,810 --> 00:43:29,240
A Ponzi scheme was named after
Charles Ponzi, a Boston

947
00:43:29,240 --> 00:43:32,770
investor, and here's how the
Ponzi scheme worked.

948
00:43:32,770 --> 00:43:33,560
I teach two classes.

949
00:43:33,560 --> 00:43:37,460
I teach now and I teach
again at 2:30.

950
00:43:37,460 --> 00:43:39,730
I teach 14.01 now,
14.41 at 2:30.

951
00:43:39,730 --> 00:43:42,810
Let's say I walked in here this
morning and I said, tell

952
00:43:42,810 --> 00:43:47,390
you what guys, you each give
me $1 and on Wednesday I'll

953
00:43:47,390 --> 00:43:49,390
give you $2.

954
00:43:49,390 --> 00:43:51,450
Now you're thinking, well,
that's bullshit, but you know

955
00:43:51,450 --> 00:43:52,765
what, I want a good grade so
I'll give him the dollar.

956
00:43:52,765 --> 00:43:54,790
So you give me the dollar.

957
00:43:54,790 --> 00:43:58,850
I then go to my 2:30 1441 class
and say, hey guys, you

958
00:43:58,850 --> 00:44:06,040
each give me $3 and on Wednesday
I'll give you $5.

959
00:44:06,040 --> 00:44:08,210
They are like same thing,
whatever, but they're higher

960
00:44:08,210 --> 00:44:09,390
classmen than you, they
have more money,

961
00:44:09,390 --> 00:44:11,340
they give me the money.

962
00:44:11,340 --> 00:44:14,560
I take their $3, I pocket $1.

963
00:44:14,560 --> 00:44:17,210
I come back on Wednesday
and give you each $2.

964
00:44:17,210 --> 00:44:18,730
You're like, holy shit,
that worked.

965
00:44:18,730 --> 00:44:20,710
I'm like, I've got a better
deal for you.

966
00:44:20,710 --> 00:44:24,610
You guys give me $6 and next
time I see you in class--

967
00:44:24,610 --> 00:44:26,160
semester's ending, but
you get the point--

968
00:44:26,160 --> 00:44:27,900
I'll give you $10.

969
00:44:27,900 --> 00:44:29,790
You're like, great, this guy's
trustworthy, we'll do that.

970
00:44:29,790 --> 00:44:30,880
You give me the $6.

971
00:44:30,880 --> 00:44:31,650
I pocket $1,

972
00:44:31,650 --> 00:44:35,080
I go back to my 14.41 and say
here's the $5 I owe you.

973
00:44:35,080 --> 00:44:36,020
They say, wow, that's great.

974
00:44:36,020 --> 00:44:37,410
I say, so I've got a better
deal for you.

975
00:44:37,410 --> 00:44:40,080
You give me $11, and
so on and so on.

976
00:44:40,080 --> 00:44:43,230
And as long as you're both
willing to play I can continue

977
00:44:43,230 --> 00:44:44,490
to pocket the money
and make money.

978
00:44:44,490 --> 00:44:46,470
This is also called the
pyramid scheme.

979
00:44:46,470 --> 00:44:49,440
You might have heard a guy
named Bernie Madoff.

980
00:44:49,440 --> 00:44:52,850
Never invest your money with a
guy who's last name is Madoff.

981
00:44:52,850 --> 00:44:57,230
Bernie Madoff, for 20 years ran
this kind of Ponzi scheme.

982
00:44:57,230 --> 00:45:00,710
What he'd do is he'd find
investors, he'd take money

983
00:45:00,710 --> 00:45:03,400
from them and promise them a
return, he'd then find more

984
00:45:03,400 --> 00:45:05,410
investors, take the money from
the new investors and pay off

985
00:45:05,410 --> 00:45:07,860
the old investors,
and keep going.

986
00:45:07,860 --> 00:45:09,570
There was nothing real behind
what he was doing.

987
00:45:09,570 --> 00:45:12,480
He made it work for 20 years
until someone didn't want to

988
00:45:12,480 --> 00:45:13,770
play anymore.

989
00:45:13,770 --> 00:45:15,945
What if we did this, and it
was the last class and you

990
00:45:15,945 --> 00:45:17,080
guys said, well, see you
Professor Gruber.

991
00:45:17,080 --> 00:45:18,910
I'm like, wait a second, if you
give me $20, and you're

992
00:45:18,910 --> 00:45:20,430
out the door.

993
00:45:20,430 --> 00:45:22,110
Now I got to go to my next class
and I don't have any

994
00:45:22,110 --> 00:45:24,080
money for them.

995
00:45:24,080 --> 00:45:27,150
Then the whole thing collapses
And it collapses because a

996
00:45:27,150 --> 00:45:30,200
Ponzi scheme is not
a real investment.

997
00:45:30,200 --> 00:45:33,060
The asset behind a Ponzi scheme
is not a bank account.

998
00:45:33,060 --> 00:45:36,920
The asset behind a Ponzi
scheme is trust. And

999
00:45:36,920 --> 00:45:38,840
basically, as long as someone's
no longer trusting

1000
00:45:38,840 --> 00:45:43,300
or can no longer pay
in, it goes away.

1001
00:45:43,300 --> 00:45:45,260
Now, here's a little secret.

1002
00:45:45,260 --> 00:45:48,070
Our Social Security system
is a Ponzi scheme.

1003
00:45:48,070 --> 00:45:49,255
The nation's largest
social insurance

1004
00:45:49,255 --> 00:45:50,420
program is a Ponzi scheme.

1005
00:45:50,420 --> 00:45:51,560
How is that so?

1006
00:45:51,560 --> 00:45:54,400
Well because you guys are
paying in now, and those

1007
00:45:54,400 --> 00:45:56,080
checks are going to
today's elders.

1008
00:45:56,080 --> 00:45:59,540
There's no guarantee that 40
years from now kids will pay

1009
00:45:59,540 --> 00:46:00,910
in to support you.

1010
00:46:00,910 --> 00:46:01,900
No guarantee at all.

1011
00:46:01,900 --> 00:46:04,800
You can pay in your whole life,
and then people could

1012
00:46:04,800 --> 00:46:06,310
vote out Social Security
when you're

1013
00:46:06,310 --> 00:46:09,350
retired and you're screwed.

1014
00:46:09,350 --> 00:46:10,160
It's a Ponzi scheme.

1015
00:46:10,160 --> 00:46:14,760
There's nothing behind this
except trust in the

1016
00:46:14,760 --> 00:46:16,450
government's programs.

1017
00:46:16,450 --> 00:46:18,630
The only asset behind this is
the trust the government would

1018
00:46:18,630 --> 00:46:21,050
never screw the voter,
the elderly voters,

1019
00:46:21,050 --> 00:46:22,830
by ending the program.

1020
00:46:22,830 --> 00:46:24,950
As long as the government is
willing to force young people

1021
00:46:24,950 --> 00:46:27,050
to pay checks to support
older people the

1022
00:46:27,050 --> 00:46:28,580
program can never end.

1023
00:46:28,580 --> 00:46:31,045
But if the government ever is
unwilling to do that then a

1024
00:46:31,045 --> 00:46:33,540
whole generation is going to
suffer from the Ponzi scheme.

1025
00:46:33,540 --> 00:46:36,260
Now that's pretty unlikely
because elderly voters are

1026
00:46:36,260 --> 00:46:37,200
pretty powerful.

1027
00:46:37,200 --> 00:46:39,370
It seems unlikely that Ponzi
scheme will ever collapse.

1028
00:46:39,370 --> 00:46:41,170
That's why we don't think of
it in those terms. But it

1029
00:46:41,170 --> 00:46:42,120
really is just that.

1030
00:46:42,120 --> 00:46:44,510
It's just a Ponzi scheme, in
this case supported by trust

1031
00:46:44,510 --> 00:46:45,930
in the government.

1032
00:46:45,930 --> 00:46:50,620
The problem is this Ponzi scheme
works well so long as

1033
00:46:50,620 --> 00:46:53,620
more people are paying in than
are collecting money.

1034
00:46:53,620 --> 00:46:59,940
So right now we have, in the US,
we have about 10 workers

1035
00:46:59,940 --> 00:47:02,560
for every elderly person, or
about eight workers for every

1036
00:47:02,560 --> 00:47:04,080
elderly person.

1037
00:47:04,080 --> 00:47:06,770
So we can collect a lot of money
to pay off the elderly.

1038
00:47:06,770 --> 00:47:09,520
The problem is that number's
diminishing rapidly because of

1039
00:47:09,520 --> 00:47:10,660
the aging of the baby boomers.

1040
00:47:10,660 --> 00:47:13,590
There's a huge bulge of
individuals who are about to

1041
00:47:13,590 --> 00:47:15,530
become elderly and collect
their Social Security.

1042
00:47:18,730 --> 00:47:24,090
By the year 2025 that 8:1 will
ratio be more like 4:1, and

1043
00:47:24,090 --> 00:47:25,130
we're suddenly going to
be short on money.

1044
00:47:25,130 --> 00:47:27,950
Here's the way I think about
it most scarily.

1045
00:47:27,950 --> 00:47:30,530
How many of you guys have
grandparents in Florida?

1046
00:47:30,530 --> 00:47:32,740
OK, you've got a few
grandparents in Florida.

1047
00:47:32,740 --> 00:47:37,020
The US in 2025, the US as a
whole, will be older than

1048
00:47:37,020 --> 00:47:37,950
Florida is today.

1049
00:47:37,950 --> 00:47:39,760
Because if you think of Florida
as full of old people,

1050
00:47:39,760 --> 00:47:42,990
that's the US in 20 years,
in 15 years.

1051
00:47:42,990 --> 00:47:45,690
We're moving to a very aging
society, and that's going to

1052
00:47:45,690 --> 00:47:47,880
put enormous pressure on
this Ponzi scheme.

1053
00:47:47,880 --> 00:47:50,600
And that's why, in about 30
years or in about 25 years,

1054
00:47:50,600 --> 00:47:53,500
Social Security goes bankrupt.

1055
00:47:53,500 --> 00:47:57,410
We actually run out of money
unless we fix the program.

1056
00:47:57,410 --> 00:47:59,740
So currently we have a situation
where we are

1057
00:47:59,740 --> 00:48:01,530
projected for that program to go
bankrupt-- it's actually in

1058
00:48:01,530 --> 00:48:02,660
about 30 years--

1059
00:48:02,660 --> 00:48:05,180
the program to go bankrupt
because we have to pay out so

1060
00:48:05,180 --> 00:48:07,910
much to this large
group of elderly.

1061
00:48:07,910 --> 00:48:09,390
And that's a problem the
government needs to deal with.

1062
00:48:09,390 --> 00:48:12,280
And it's going to have to deal
with it in one of two ways.

1063
00:48:12,280 --> 00:48:15,340
It's going to either have
to raise taxes.

1064
00:48:15,340 --> 00:48:17,810
We currently pay about
12.4% in payroll.

1065
00:48:17,810 --> 00:48:20,080
To solve this problem forever
that tax rate would have to go

1066
00:48:20,080 --> 00:48:22,530
up by about three and
a half to 4%.

1067
00:48:22,530 --> 00:48:24,320
So we'd have to pay the three
and a half to 4% more of our

1068
00:48:24,320 --> 00:48:26,570
wages to solve this problem.

1069
00:48:26,570 --> 00:48:29,490
Or we're going to have to cut
benefits, do things like

1070
00:48:29,490 --> 00:48:31,460
saying to people they can't get
their benefits at 62, that

1071
00:48:31,460 --> 00:48:33,220
goes up to 64.

1072
00:48:33,220 --> 00:48:35,180
And you can see US style rioting
rather than French

1073
00:48:35,180 --> 00:48:37,630
style rioting.

1074
00:48:37,630 --> 00:48:39,800
That's the difficult problem
the government faces going

1075
00:48:39,800 --> 00:48:42,170
forward is that something
has to give.

1076
00:48:42,170 --> 00:48:45,490
We either have to raise taxes to
support this program or we

1077
00:48:45,490 --> 00:48:47,910
have to cut the benefits that
elderly people get.

1078
00:48:47,910 --> 00:48:49,260
Yeah?

1079
00:48:49,260 --> 00:48:52,427
AUDIENCE: This might just
postpone the problem, but

1080
00:48:52,427 --> 00:48:54,784
would it be possible for
the government to

1081
00:48:54,784 --> 00:48:56,748
taper off the benefits?

1082
00:48:56,748 --> 00:49:00,760
So for instance, at 62, when
you're eligible, you could get

1083
00:49:00,760 --> 00:49:03,773
let's say 50% of what you'd
normally would be getting, and

1084
00:49:03,773 --> 00:49:06,570
that amount would slowly
increase as you get older.

1085
00:49:06,570 --> 00:49:09,850
So when you're making your
labor decisions you would

1086
00:49:09,850 --> 00:49:13,562
work, but maybe only half or
quarter time what you normally

1087
00:49:13,562 --> 00:49:14,954
would work and that might--

1088
00:49:14,954 --> 00:49:17,420
PROFESSOR: Certainly there are
compromises along the way but

1089
00:49:17,420 --> 00:49:19,130
that still goes in the category
of benefit cuts.

1090
00:49:19,130 --> 00:49:20,220
You're just talking about
a different way to

1091
00:49:20,220 --> 00:49:21,040
do a benefits cut.

1092
00:49:21,040 --> 00:49:23,390
There's lots of very subtle
and interesting ways--

1093
00:49:23,390 --> 00:49:25,280
and believe me, when the time
comes I'm not just going to

1094
00:49:25,280 --> 00:49:27,190
say we're cutting everybody's
benefit 10%.

1095
00:49:27,190 --> 00:49:28,920
It will be something much more
subtle along the lines you're

1096
00:49:28,920 --> 00:49:31,050
describing, sufficiently
complicated that people don't

1097
00:49:31,050 --> 00:49:34,400
understand it so that they can
try and slip it by people.

1098
00:49:34,400 --> 00:49:35,510
If they're going to cut benefits
it's going to be that

1099
00:49:35,510 --> 00:49:38,640
complicated, and that's the
trade off we're going to face.

1100
00:49:38,640 --> 00:49:38,950
All right.

1101
00:49:38,950 --> 00:49:39,190
OK.

1102
00:49:39,190 --> 00:49:43,090
So next class we'll talk about
health insurance and we'll do

1103
00:49:43,090 --> 00:49:44,340
a final session.