1
00:00:00,000 --> 00:00:04,480
The following is part of Cornell 
Contemporary China Initiative lecture series  

2
00:00:04,480 --> 00:00:08,240
under the Cornell East Asia Program. 
The arguments and viewpoints of this  

3
00:00:08,240 --> 00:00:11,600
talk belong solely to the 
speaker. We hope you enjoy.

4
00:00:11,600 --> 00:00:18,160
Uh so today we have Ruixue Jia 
come to speak to us from San Diego,  

5
00:00:18,160 --> 00:00:24,960
where she's been since 2013, uh but before 
that she was in Stockholm where she did her PhD  

6
00:00:26,560 --> 00:00:30,880
in, and I have to read this too because 
it's long to remember, in the Institute  

7
00:00:30,880 --> 00:00:36,400
for International Economic Studies at Stockholm 
University, uh and your department now in San  

8
00:00:36,400 --> 00:00:46,640
Diego is the Global Policy and Strategy, GPS, 
newly named. Uh so we're very happy to have her  

9
00:00:46,640 --> 00:00:51,600
here from all these uh disparate places to 
speak to us about some very important things  

10
00:00:51,600 --> 00:00:57,520
in China uh very interesting mix of economics 
and social issues, so my hand it to you.

11
00:00:57,520 --> 00:01:06,560
Thank you uh thanks a lot uh so this is a based on 
mainly a joint work with Hongbin Li from Tsinghua,  

12
00:01:06,560 --> 00:01:12,400
so you see his name there. Let's before 
talking about China and the exam, I want to  

13
00:01:12,960 --> 00:01:17,200
share with you a broad motivation, why 
we started thinking about this project,  

14
00:01:18,480 --> 00:01:24,480
Access to education, especially elite 
education, is believed to be super important  

15
00:01:24,480 --> 00:01:29,840
for elite formation and social mobility 
across all societies in modern societies,  

16
00:01:30,480 --> 00:01:34,640
You may know the sociological works 
by Pierre Bourdieu, for example,  

17
00:01:35,280 --> 00:01:40,640
and I will not talk about that literature, 
instead of more, talk more about economics. So  

18
00:01:40,640 --> 00:01:47,680
in economics, there's also burgeoning literature, 
trying to estimate the returns to elite education.  

19
00:01:47,680 --> 00:01:53,440
For example, if you attended university, Cornell 
university how much more would you earn after  

20
00:01:53,440 --> 00:02:00,720
graduation compared to attending a non-elite or 
non-Ivy League university? Surprisingly or not,  

21
00:02:01,440 --> 00:02:08,160
there's no conclusion on this, and many studies 
find actually you don't necessarily earn more, why  

22
00:02:08,160 --> 00:02:14,000
there's no, why there's controversy, conversation 
about this topic, you, you would think, oh it may  

23
00:02:14,000 --> 00:02:19,920
be easier just compare the graduate students here 
with another set of graduate students from a less  

24
00:02:19,920 --> 00:02:25,200
good university. Let's look at their salaries and 
we get the conclusion, but that's not easy. As  

25
00:02:25,200 --> 00:02:32,240
you see, uh you know, your, you can come to Cornell 
because you are much able than other universities,  

26
00:02:32,240 --> 00:02:38,640
or your family background might be very different 
that students from non-elite universities, so  

27
00:02:38,640 --> 00:02:43,120
in the end when you go to the job market even 
if you earn more it's not necessarily because  

28
00:02:43,120 --> 00:02:47,760
of Cornell, it could be because of your own 
ability or because your family background  

29
00:02:47,760 --> 00:02:52,480
so that's a typical challenge known as the 
selection, and that's, you know, makes it  

30
00:02:52,480 --> 00:02:57,920
difficult to reach a conclusion about whether 
attending elite university really helps or not  

31
00:02:58,640 --> 00:03:04,160
and this literature also didn't talk much 
about elite formation, social mobility.  

32
00:03:04,160 --> 00:03:10,160
So in this project, we try to first also 
estimate the returns to elite education,  

33
00:03:10,160 --> 00:03:16,240
but we also want to speak to social mobility and 
elite formation in China. In addition, we want to  

34
00:03:16,240 --> 00:03:22,080
shed some light to understand the mechanism. If 
we find a wage premium, where does it come from,  

35
00:03:23,680 --> 00:03:29,120
and why China, of course? My reason is: I'm 
Chinese. I care about China, and many of you  

36
00:03:29,120 --> 00:03:33,920
may care about China; that's why you are here. 
But there are even, if you you know don't care  

37
00:03:33,920 --> 00:03:40,240
about China, a Chinese case is very interesting 
to study intellectually because uh this Chinese  

38
00:03:40,240 --> 00:03:46,400
ideal laboratory to understand the rule of 
elite education. As you may know, each year,  

39
00:03:46,400 --> 00:03:52,880
around 10 million students are taking this college 
entrance exam. This is likely to be the largest  

40
00:03:52,880 --> 00:03:59,040
standardized exam in the world, and this exam 
score will determine whether a young person can  

41
00:03:59,040 --> 00:04:06,880
attend a university, and which tier of university, 
so it's really important for this million,  

42
00:04:06,880 --> 00:04:13,440
10 million of students each year, and they exist 
cut-offs for different tiers of university,  

43
00:04:13,440 --> 00:04:17,920
so if you are just, for the first tier, the 
elite university, if you're above the cut-off,  

44
00:04:17,920 --> 00:04:23,200
you're eligible to apply, if not you you're not 
eligible. When social scientists hear about this  

45
00:04:23,200 --> 00:04:29,600
they get excited. You seem not very excited, but 
this, because, why this is so exciting because now  

46
00:04:29,600 --> 00:04:34,400
you can compare relatively similar students, 
you know, before I talk, I said it's very  

47
00:04:34,400 --> 00:04:41,120
difficult to compare students in Cornell with the 
non-Ivy League university students, but now you  

48
00:04:41,120 --> 00:04:46,640
can compare relatively similar students, it just 
someone are just a few points above the cut off,  

49
00:04:46,640 --> 00:04:52,160
some are just few points below the cutoff. Now 
let's look at which university they went to,  

50
00:04:52,160 --> 00:04:58,400
and let's follow them, see whether they're someone 
who, uh whether there's difference in which when  

51
00:04:58,400 --> 00:05:04,800
they go to the labor market. And in addition to 
this general motivation, the exam system may be  

52
00:05:04,800 --> 00:05:10,560
one of the most important institutions in China, 
as many of you may know. It's often perceived to  

53
00:05:10,560 --> 00:05:16,880
be, you know, the only most important channel 
for Communists to become elite in the society.  

54
00:05:17,600 --> 00:05:24,080
I will focus on economic consequence in this talk, 
but this exam system also has other interesting  

55
00:05:24,080 --> 00:05:28,160
social, cultural, political consequence. 
If you actually if you are interested in  

56
00:05:28,160 --> 00:05:34,240
the political consequence you can read a great 
paper, I was joking, written by my co-author and I  

57
00:05:34,880 --> 00:05:40,880
last year, look at a historical version of the 
exam, and how that affects political stability  

58
00:05:40,880 --> 00:05:45,360
of Chinese dynasties, but I wouldn't talk 
about this in this talk. I'll talk about  

59
00:05:45,360 --> 00:05:52,560
the modern college entrance exam and the economic 
consequences. Just to show you some picture, uh,  

60
00:05:52,560 --> 00:05:58,320
you know you can easily, if you type the exam 
called gao kao (高考) in Chinese in Google,  

61
00:05:58,320 --> 00:06:04,480
you'll find many many impressive pictures online. 
Uh the first pictures, I select a few for you,  

62
00:06:04,480 --> 00:06:10,560
the first one is just like how my classroom 
looked like when I was in high school. You know,  

63
00:06:10,560 --> 00:06:16,320
people are just heading down, practicing the the 
exercises for the exam, they don't have time to  

64
00:06:16,320 --> 00:06:21,360
ask questions. So that's one, I always 
think that's one reason why Chinese students  

65
00:06:21,360 --> 00:06:26,240
are not great at asking questions, even when 
they come to the U.S., so I'm very happy I'm  

66
00:06:26,240 --> 00:06:30,640
on this side speaking instead of sitting there 
thinking about oh what questions should I ask,  

67
00:06:31,200 --> 00:06:37,520
so don't worry if you do not have many 
questions later. And the second picture uh is a  

68
00:06:38,240 --> 00:06:42,640
not only the students are clearly working 
hard, their parents are invest a lot,  

69
00:06:42,640 --> 00:06:47,680
and when they were like writing on the exam during 
like the two days in June, their parents are  

70
00:06:47,680 --> 00:06:53,120
waiting equally anxiously outside. You know, my 
father was doing that when I was taking the exam.  

71
00:06:53,840 --> 00:06:59,360
And the schools also came up all type interesting 
slogans trying to encourage their students to work  

72
00:06:59,360 --> 00:07:05,440
hard, and here's one popular one, said: oh without 
the exam, how could you compete with the second  

73
00:07:05,440 --> 00:07:11,840
generation of the rich, you know. This this is 
because people some students complain that it  

74
00:07:11,840 --> 00:07:17,440
is so harsh to work for the exam but the 
rationale is, without it how could you  

75
00:07:17,440 --> 00:07:23,040
compete in this society, because the children 
of the rich has so much, so many advantages.  

76
00:07:24,560 --> 00:07:30,320
But now, why do we, I can stop here and let you 
go. Oh this is very important, the students are  

77
00:07:30,320 --> 00:07:35,280
work hard in, parents are invest a lot, then of 
course on the labor market there should be some  

78
00:07:35,280 --> 00:07:41,520
consequence and we can stop and you can go home, 
but this is not without controversy, because this  

79
00:07:41,520 --> 00:07:47,600
in, this China, as you know family background is 
also believed to be crucial in the labor market,  

80
00:07:47,600 --> 00:07:53,840
and there was even phrase, ironical phrase 
called: to each according to his dead. You know,  

81
00:07:53,840 --> 00:08:00,320
Marxism has this to each according to his need. 
Then this is ironically say, oh in the end,  

82
00:08:00,320 --> 00:08:06,000
it's to each the distribution of the societies 
according to your debt, so if this is right,  

83
00:08:06,000 --> 00:08:11,840
then this upper mobility, upward mobility provided 
by the exam might be just an illusion rather than  

84
00:08:11,840 --> 00:08:17,120
reality. So that's why we want to collect 
systematic data and doing careful analysis  

85
00:08:17,120 --> 00:08:21,680
to better understand the general pattern, and if 
we find something, we also want to understand the  

86
00:08:21,680 --> 00:08:29,760
mechanism behind the pattern. So I just give you a 
preview about what we find so far, uh then you can  

87
00:08:29,760 --> 00:08:35,280
you know think about your Halloween party while 
I go to the details. So the first is, you know,  

88
00:08:35,280 --> 00:08:42,720
that's scoring the part, the elite university 
cut-off, make one more likely to go to university.  

89
00:08:42,720 --> 00:08:48,240
If you think, you know, China is super corrupt, 
maybe the education system is also very corrupt,  

90
00:08:48,240 --> 00:08:52,480
even there are some cutoffs, it doesn't 
matter in the end. But the answer is,  

91
00:08:53,040 --> 00:08:58,720
yes it matters a lot, you know, scoring above 
the cut-off increase the probability a lot  

92
00:08:58,720 --> 00:09:05,120
of going to elite university. And the second, the 
question is, or, does elite education lead to a  

93
00:09:05,120 --> 00:09:10,800
high wage. The answer is yes, and they decide, 
it's the effective size of what I'll show you,  

94
00:09:10,800 --> 00:09:16,800
and then you if you earn a high wage due to elite 
education, does that mean that you become elite  

95
00:09:16,800 --> 00:09:23,200
in the society? Our answer is unclear. I 
will be more concrete about this later,  

96
00:09:23,200 --> 00:09:28,880
so while we find a wage premium, we don't 
find much in other important dimension that's  

97
00:09:28,880 --> 00:09:34,080
important for being an elite in this society 
such as perks, such as whether you can working  

98
00:09:34,080 --> 00:09:40,240
for the government, whether you can work for the 
banking industry, etc. I'll be more uh clear later  

99
00:09:40,240 --> 00:09:46,320
uh and we also want to know that's going to, the 
in terms of intergenerational mobility, like,  

100
00:09:46,320 --> 00:09:51,520
that's going to, go into elite university 
or scoring the culture above the cut off,  

101
00:09:51,520 --> 00:09:56,800
changing the intergenerational mobility. In other 
words, does that change how your parental status  

102
00:09:57,680 --> 00:10:03,840
affect your own status? And our answer is: 
no. This might be surprising, but we find no  

103
00:10:04,480 --> 00:10:10,080
change in inter-generational mobility. I'll be 
clear later about what we find. Finally, we want  

104
00:10:10,080 --> 00:10:16,400
to understand what leads to the wage premium. 
Is that you learn more in an elite university  

105
00:10:16,400 --> 00:10:22,320
like Cornell, or you actually get to know more 
people so that you have a better job or that,  

106
00:10:22,320 --> 00:10:28,160
this is purely a signaling channel, you don't 
necessarily learn more, have more human capital  

107
00:10:28,160 --> 00:10:33,520
than other students in a non-elite university, 
but you still get high wage because there's strong  

108
00:10:33,520 --> 00:10:38,560
signaling, you have an elite education, that's 
very important. And what we find surprising or  

109
00:10:38,560 --> 00:10:45,040
not is more evidence for signaling than human 
capital or social networks. And you know,  

110
00:10:45,040 --> 00:10:52,240
you might have some questions and we can discuss 
later. So I will now come to the details,  

111
00:10:52,240 --> 00:10:58,400
uh I will first briefly introduce some background 
of the how elite education works in China, and  

112
00:10:58,400 --> 00:11:04,640
tells you how we collect this data. And then I'll 
go to the questions and our answers step by step.  

113
00:11:06,480 --> 00:11:12,240
So all tier, first, all tiers of Chinese 
universities, all the university in China,  

114
00:11:12,240 --> 00:11:17,520
recruit based on the exam system, and what we 
call elite university in this project is the first  

115
00:11:17,520 --> 00:11:26,240
tier. And they recruit first after the exam. They 
thee are about around, about 2,300 universities  

116
00:11:26,240 --> 00:11:34,160
in China among the (unintelligible) study period only 
96 count as the first tier across all provinces,  

117
00:11:34,160 --> 00:11:40,160
and this definition is very closely related 
to this Project 211, I'm not sure whether  

118
00:11:40,160 --> 00:11:48,240
you heard about it, but this is a pro-, this is a 
designated 100 universities for the 21st century,  

119
00:11:48,240 --> 00:11:52,640
but these are another category of elite 
universities they heavily overlapped with  

120
00:11:52,640 --> 00:11:58,800
what we call elite university in this project. And 
just for your information, all elite university in  

121
00:11:58,800 --> 00:12:05,440
China are public, not not like Cornell, I guess, 
and they are not so small they're more like UC,  

122
00:12:05,440 --> 00:12:11,120
University of California type of university and 
they don't charge higher tuition fees than private  

123
00:12:11,120 --> 00:12:18,400
universities, and the yearly tuition fee is about 
5000 RMB, so it's about 800 dollars. Not too high,  

124
00:12:18,400 --> 00:12:25,840
right, 800 dollars. Uh and what makes elite 
university special is they have more resources  

125
00:12:25,840 --> 00:12:32,720
and the government uh government support, and 
they attract very different uh students. Uh and  

126
00:12:32,720 --> 00:12:37,520
there's an interesting feature, I'll come back 
to this after telling you all those findings,  

127
00:12:37,520 --> 00:12:44,400
the education system in China is is very well 
known for strict interest and easy out. So you  

128
00:12:44,400 --> 00:12:49,680
have it has this very strict exam to select 
students, but once you are in almost everyone  

129
00:12:49,680 --> 00:12:55,120
are promised to graduate to graduate. So that 
might affect, you know, the mechanism of wage  

130
00:12:55,120 --> 00:13:00,240
premium you'll find later. So so you know if you 
go to elite university, or non-elite university,  

131
00:13:00,240 --> 00:13:07,760
the graduate race, ratio, or rate is very similar. 
Uh so how would this do, how does university  

132
00:13:08,720 --> 00:13:13,600
admit students? So just give you a 
rough picture, before most provinces  

133
00:13:13,600 --> 00:13:19,840
the process works as follows. First everyone 
takes the the exam, and the exam was graded by  

134
00:13:20,480 --> 00:13:27,920
province, so the score are only comparable 
within the province, and after seeing the scores  

135
00:13:27,920 --> 00:13:33,520
the government will decide the cutoff for the 
elite university, and the cutoff would be public,  

136
00:13:33,520 --> 00:13:38,640
public everyone knows this, and the students know 
their scores, and they know the cutoff, they fill  

137
00:13:38,640 --> 00:13:45,840
in the application, including the university and 
the majors, and if you are above the cut off,  

138
00:13:45,840 --> 00:13:51,440
the first tier cutoff, you are eligible to apply 
for these elite universities, but you are not,  

139
00:13:51,440 --> 00:13:57,040
this is not sufficient, you are not, for example, 
if you are you like 20 points above the cutoff,  

140
00:13:57,040 --> 00:14:02,000
you apply to Tsinghua, which is the top elite 
university, you are very less likely to get  

141
00:14:02,000 --> 00:14:08,080
into because many people apply in to Tsinghua, 
so there's computation in the end, uh whether  

142
00:14:08,080 --> 00:14:12,880
you know the minimum score for each university 
is different, depending on the computation.  

143
00:14:12,880 --> 00:14:17,920
So the cutoff determines the eligibility, 
but it's not sufficient, just to be clear.  

144
00:14:19,280 --> 00:14:27,440
And then after receiving these applications, 
the university decides whom to accept. So just  

145
00:14:27,440 --> 00:14:37,520
how the scores are the look like, so in this 
period most provinces use a scale of 750 points,  

146
00:14:37,520 --> 00:14:43,360
so they're based on four projects, there are two 
tracks: one is called the natural science track,  

147
00:14:43,360 --> 00:14:49,280
the other is social science track, and they are 
testing different subjects. I, I wouldn't read it,  

148
00:14:49,280 --> 00:14:56,000
but roughly like based on four subjects and 
together the maximum score is 750 points.  

149
00:14:57,520 --> 00:15:03,040
And I just want to be clear about the 
background. If you are below the cutoff,  

150
00:15:03,040 --> 00:15:09,280
you have a very low probability, not zero, but low 
probability. You can still apply if you have some  

151
00:15:09,280 --> 00:15:15,360
extra scores, for example, these extra scores 
are typically because you're a minority, or you  

152
00:15:15,360 --> 00:15:22,640
your father died for the country, or you get some 
international prize from math uh physics type of  

153
00:15:22,640 --> 00:15:27,600
competition, you have got some extra score, so 
even if your original score is below the cutoff,  

154
00:15:27,600 --> 00:15:32,240
with extra score you can become eligible, 
but the probability is generally very low  

155
00:15:32,240 --> 00:15:37,120
and if you are above the cutoff, yes you 
can apply now, but as I said, before it's  

156
00:15:37,120 --> 00:15:42,880
not that 100 sure you're getting to the at the 
university you apply, you could be rejected,  

157
00:15:42,880 --> 00:15:49,440
and didn't get get into any university. This 
is the background now, come to the data, like  

158
00:15:49,440 --> 00:15:53,920
I said and the social scientists are very excited 
about this, they should have already done this,  

159
00:15:53,920 --> 00:16:00,240
why wouldn't, why would no one has done this, this 
is because there's no data on this. Individuals do  

160
00:16:00,240 --> 00:16:06,400
know their own score but there's no systematic 
score available for large group of individual,  

161
00:16:06,400 --> 00:16:12,560
even you get that it's very difficult to 
link the scores to the labor market outcome,  

162
00:16:12,560 --> 00:16:18,320
so to do this one has to collect data myself. So 
what we do is we have been collecting this data  

163
00:16:18,960 --> 00:16:29,120
for six years by surveying the graduating cohort in 
May and June during 2010 and 2015. So as you may  

164
00:16:29,120 --> 00:16:36,000
know, students would already, they would have to 
leave their school or their college in July by May  

165
00:16:36,000 --> 00:16:41,760
and June, they already know what they would do for 
the future, right, some a small share would go to  

166
00:16:41,760 --> 00:16:46,960
graduate school, but the majority of them already 
searched for job, and then the majority of them  

167
00:16:46,960 --> 00:16:53,760
have already got job offers, and that's what we 
get about their job outcomes. And of course this  

168
00:16:53,760 --> 00:16:58,960
is the first job. What I show you, the main result 
of the first job but I'll show you some additional  

169
00:16:58,960 --> 00:17:05,120
(unintelligible) time additional data telling you that 
the first job is very important for your future  

170
00:17:05,120 --> 00:17:11,680
jobs too, and this might be useful for you since 
you have not searched for a job yet, I guess,  

171
00:17:11,680 --> 00:17:17,920
so actually yesterday I was in a conference on 
inequality in the U.S. People using detailed  

172
00:17:17,920 --> 00:17:24,400
data in this country showing that the inequality 
within a cohort, so with, when the graduating  

173
00:17:24,400 --> 00:17:32,080
cohort in lifetime income is already determined 
at age 20- 25 so basically your first job is  

174
00:17:32,080 --> 00:17:38,080
super important in determining your lifetime 
income. And this is also the case in China,  

175
00:17:39,360 --> 00:17:46,480
I wouldn't have time to tell you how how 
we implemented these surveys step by step,  

176
00:17:46,480 --> 00:17:53,040
but one interesting thing is we have very high 
response rate. This is because we not really give  

177
00:17:53,040 --> 00:18:00,480
the survey questionnaires by individuals, instead 
we selected all the individuals by some random  

178
00:18:00,480 --> 00:18:06,800
sample method and then we gathered all of them in 
this such a big classroom, and they fill all the  

179
00:18:06,800 --> 00:18:13,840
questionnaires by the same day, the same time, 
the same place that increase the response rate.  

180
00:18:14,880 --> 00:18:22,160
So for, for these six years we've managed to 
survey exactly 90 universities in China. Among  

181
00:18:22,160 --> 00:18:31,040
them, 26 belong to the amount of of the elite 
university group. We deliberately oversampled  

182
00:18:31,040 --> 00:18:38,000
the elite university so that we can have enough 
students to compare. So in these six years,  

183
00:18:38,000 --> 00:18:46,080
we collected information for about 4,000, 40,000 
uh students in the in their graduating year,  

184
00:18:46,080 --> 00:18:51,120
so the first map, just to show you where the 
ninety universities were located, were located,  

185
00:18:51,120 --> 00:18:58,000
there as you see they're widely spread across 26 
provinces in China, and if you look at where their  

186
00:18:58,000 --> 00:19:03,920
students come from, they come from all over China, 
you know, we don't have like university in Tibet,  

187
00:19:04,800 --> 00:19:11,920
but you know we still have students from Tibet in 
other provinces, so we have about 40,000 students  

188
00:19:11,920 --> 00:19:18,720
but before today's talk I will focus on about 
10,000 students of them who have scores just  

189
00:19:18,720 --> 00:19:26,320
within 20 points, around the cut-off. If you 
recall, that the maximum score is 750 points,  

190
00:19:26,320 --> 00:19:31,920
so 20 points you know it's not that that 
large, so people are relatively comparable,  

191
00:19:31,920 --> 00:19:37,840
and so we'll compare students all of them 
are within 20 points around the cut-off.

192
00:19:40,080 --> 00:19:48,400
So now I come to the first set result to show 
you that how the exam scores affect access to  

193
00:19:48,400 --> 00:19:53,440
elite education, and also show you a little 
bit about how this affects other dimensions,  

194
00:19:53,440 --> 00:20:01,200
like what major they would be majoring, where they 
study, and their rank in college. So the first uh  

195
00:20:01,840 --> 00:20:09,520
figure, the x-axis is is the point to the 
cutoff, so each dot indicate or on one point,  

196
00:20:09,520 --> 00:20:15,680
not um a group of students are one point below, 
uh or five points below the cutoff or five points  

197
00:20:15,680 --> 00:20:22,720
above the cutoff. The y axis plots the average 
probability of attending an elite university,  

198
00:20:22,720 --> 00:20:28,400
so what you see is, this is zero here, it's 
there's very small probability that one can  

199
00:20:28,400 --> 00:20:34,080
go to elite university if they are below 
the cutoff. But you see this very clear  

200
00:20:34,080 --> 00:20:40,880
jump after one is above the cutoff, and this 
is the increase as you're getting high scores.  

201
00:20:41,760 --> 00:20:46,640
And this is, you know, clearly show this 
significant change if you're above the cutoff,  

202
00:20:47,280 --> 00:20:53,360
you're much more likely to attend an elite 
university, as I said, because why do I put it  

203
00:20:53,360 --> 00:20:59,600
here because the scores are only comparable within 
province, year, and track. There are two tracks,  

204
00:20:59,600 --> 00:21:06,240
so sometimes you want to isolate these facts to 
just compare students within the province year and  

205
00:21:06,240 --> 00:21:12,560
track. So I only compare students within Beijing, 
just one student in Beijing with another student  

206
00:21:12,560 --> 00:21:17,440
in Beijing. I'm not to compare a student in 
Beijing with a student in Shandong, for example.  

207
00:21:18,640 --> 00:21:24,080
And you see, now after considering this, you see a 
similar pattern. So in terms of magnitude, if you  

208
00:21:24,080 --> 00:21:29,680
are scoring above the cut off, the probability, 
you, the probability of getting elite,  

209
00:21:29,680 --> 00:21:36,560
into elite university would be increased by 0.15 
but this is about, over 70 percent of the mean.  

210
00:21:36,560 --> 00:21:43,840
So your probability of going to elite university 
increases by 70 percent, and this is pretty large.  

211
00:21:44,640 --> 00:21:49,200
And and you would wonder, you know, I will 
tell you that one, why we want to use this  

212
00:21:49,200 --> 00:21:55,280
kind of design is to to make sure that the 
students around the cutoff are comparable,  

213
00:21:55,280 --> 00:22:00,640
right, so we want to check just to show you, you 
know, if you look at their well their gender,  

214
00:22:00,640 --> 00:22:07,360
their age, whether they have rural or urban, or 
their family income, you'll find that they're  

215
00:22:07,360 --> 00:22:13,280
similar, there's no similar diff- like 
this jump in these figures like implying  

216
00:22:13,280 --> 00:22:18,640
that they're fairly comparable. If you don't do 
this way, if you just look at all the students  

217
00:22:18,640 --> 00:22:25,520
then you'll also see them seem like discontinue 
you'll see a jump because students are being elite  

218
00:22:25,520 --> 00:22:30,320
university, or have high scores that typically 
also have high, from high income families,  

219
00:22:30,320 --> 00:22:36,560
etc. And this way, we can make sure that these 
students are comparable. I want to mention that,  

220
00:22:37,520 --> 00:22:42,960
it's not a, scoring about the cut off not only 
changes whether you go to an elite university,  

221
00:22:42,960 --> 00:22:49,440
it also changes your major, and it changes 
where you go to, uh you know, where you study,  

222
00:22:49,440 --> 00:22:55,520
and it changes your ranking in college. So 
how, why this, uh how does this work? So uh  

223
00:22:56,160 --> 00:23:01,680
if you think about this, if you are just above 
the cutoff, right, you are lucky about the cutoff,  

224
00:23:01,680 --> 00:23:08,080
but you are the worst in this eligible group. All 
the other students uh have higher score than you,  

225
00:23:08,080 --> 00:23:15,440
right. Is this clear? Uh so you are worse so you 
are actually less likely to go to a good major,  

226
00:23:15,440 --> 00:23:20,800
like in this period, economics, finance and 
the law are the more popular major. You're not,  

227
00:23:20,800 --> 00:23:25,920
you're less likely to be able to enroll in such 
popular major because, among those are eligible  

228
00:23:25,920 --> 00:23:32,160
for elite university, you are the worst. Uh 
whereas if you are in a non-elite group, even  

229
00:23:32,160 --> 00:23:37,840
though you are not eligible to elite university, 
among compared with your peers, you are the best.  

230
00:23:37,840 --> 00:23:42,800
So there's an interesting reversal pattern here, 
and this is what we find if you look at the  

231
00:23:42,800 --> 00:23:49,360
probability, this is the same, like score point 
by point relative to the cutoff, if you look at  

232
00:23:49,360 --> 00:23:56,960
the probability of enrolling uh econ management 
or law, you'll see actually there's, if you are  

233
00:23:56,960 --> 00:24:03,200
above the cutoff, you are slightly less likely to 
maintain these popular major affairs. And this is  

234
00:24:03,200 --> 00:24:09,120
can be explained by a slightly increasing 
STEM and slightly increase in humanity.  

235
00:24:10,400 --> 00:24:16,640
And if you look at where the university located, 
they're more likely to be outside the home  

236
00:24:16,640 --> 00:24:22,800
province, because all this elite university 
typically including be in provinces far away  

237
00:24:22,800 --> 00:24:28,160
from the home province, so they're also more 
likely to study outside their home province,  

238
00:24:29,040 --> 00:24:35,840
and as I said, they are relatively worse 
within the eligible group. This could be  

239
00:24:35,840 --> 00:24:41,600
reflected by their ranking in college, so in 
the survey we asked how do you think about  

240
00:24:41,600 --> 00:24:46,960
your rank in your class, like do you think 
you belong to 20 percent, top twenty percent  

241
00:24:46,960 --> 00:24:53,120
or bottom 20 percent? What's interesting is, 
there's a very interesting bias. For example,  

242
00:24:53,120 --> 00:25:01,440
only like only 5 percent of students claim that 
they are bottom 20. This doesn't like you know  

243
00:25:01,440 --> 00:25:07,280
makes sense, right? But that's our bias
and over 40 percent of students claim they are  

244
00:25:07,280 --> 00:25:13,840
top 20. So, we are always overconfident, right; we 
are less likely to think we are like the bottom;  

245
00:25:13,840 --> 00:25:18,880
we're more likely to think we're at the top. 
But despite this, you know, interesting bias,  

246
00:25:18,880 --> 00:25:24,800
you'll see that if you are above the cutoff, you 
are more, less likely to see you are on the top;  

247
00:25:24,800 --> 00:25:29,520
you're more likely to see you are bottom. All 
this just to to paint to the picture that,  

248
00:25:29,520 --> 00:25:34,000
yes, these people are eligible to elite 
university if they're above the cutoff,  

249
00:25:34,000 --> 00:25:39,200
but they also have some disadvantage in major 
and maybe college ranking, or this might affect  

250
00:25:39,200 --> 00:25:44,560
their labor market outcome, too, and we need 
to consider this when we come to the wage part.  

251
00:25:45,280 --> 00:25:52,480
Now I'll come to the wage part, uh I will first 
show you you know how this affects your wage, you  

252
00:25:52,480 --> 00:25:58,880
know, the first job wage and I'll need to tell you 
a little bit about how major and other dimensions  

253
00:25:58,880 --> 00:26:05,520
affect these findings, and what's the implication 
on elite status for the students and how important  

254
00:26:05,520 --> 00:26:11,440
the first job for your future jobs. Just, before 
showing the wage, I want to, you know, as you can  

255
00:26:11,440 --> 00:26:16,640
imagine, not everyone started looking for a job, 
right; some of them would go to graduate school,  

256
00:26:16,640 --> 00:26:23,120
so average 74 percent searched for a job. Among 
those searched for a job, 74 got offered, this  

257
00:26:23,120 --> 00:26:28,560
is not typo, it just happened to be the case. 
And so, so I told you in the first part,  

258
00:26:28,560 --> 00:26:34,480
about ten thousand students, we look 
at in terms of uh entering- entry uh  

259
00:26:35,680 --> 00:26:42,240
into college, and now we are looking at the 
wage part; we only have half of these students  

260
00:26:42,240 --> 00:26:48,400
really have wage information. And just for 
your information, those above the cutoff are  

261
00:26:48,400 --> 00:26:53,920
slightly less likely to look for jobs, as you 
can imagine. If you go to elite university, you  

262
00:26:53,920 --> 00:26:58,960
might have other options, et cetera, that affects 
the probability, but the difference is not huge.  

263
00:27:00,240 --> 00:27:06,960
So if you look at the wage, you you can look at 
this- the similar graph point by point, uh and  

264
00:27:06,960 --> 00:27:13,920
this is a the better way of looking at that. The 
y-axis now is the log wage if you think about  

265
00:27:13,920 --> 00:27:18,880
the difference between the, the at- the right hand 
side and the left-hand side and the right-hand  

266
00:27:18,880 --> 00:27:25,040
side, you can think of the percentage change 
in wage level. So if you look at the difference  

267
00:27:25,040 --> 00:27:30,400
between these two, you can't see clearly, so you 
have to trust me. It's about six percent, so the  

268
00:27:30,400 --> 00:27:35,600
wage difference would be about six percent. If 
you are just above the cut off, your your monthly  

269
00:27:35,600 --> 00:27:44,480
wages increased by six percent. And this is not, 
no this is about means, about uh 160 RMB is about  

270
00:27:44,480 --> 00:27:52,640
25, 26 US dollars, maybe not huge but if you look 
at wage premium, you have to divide it- if- this  

271
00:27:53,200 --> 00:28:00,320
wage difference by the probability increased 
the probability of going to elite university  

272
00:28:00,320 --> 00:28:05,920
increased by being above the cutoff. And this 
would give us an estimate about 40 percent,  

273
00:28:05,920 --> 00:28:11,600
so- which is pretty large. It means that elite 
education has a wage premium about 40 percent,  

274
00:28:11,600 --> 00:28:20,160
that about 1000 RMB in a month. In this period, I 
forgot to see if the monthly wage for these fresh  

275
00:28:20,160 --> 00:28:28,880
graduates is about 2500 RMB. So 1000 basically 
is about 40 percent of the monthly increase. So  

276
00:28:28,880 --> 00:28:34,960
roughly, you can think going to a elite university 
at the median wage which would be about 3,500.  

277
00:28:36,720 --> 00:28:43,360
There's no good like very well identified wage 
premium for college students in China, either,  

278
00:28:43,360 --> 00:28:48,320
but we can compare with them. If you look 
at other contests, uh like in the U.S.,  

279
00:28:48,320 --> 00:28:54,400
some estimate is zero, so the elite university 
like the Kruger-Dale study there's no  

280
00:28:54,400 --> 00:29:01,600
elite education wage premium. There's one study 
in Italy compared students who could just go to  

281
00:29:02,320 --> 00:29:06,720
Bocconi versus other university and 
they had a estimate of 45 percent,  

282
00:29:07,280 --> 00:29:14,480
but it's difficult to compare across countries. I 
haven't considered the major or like location of  

283
00:29:14,480 --> 00:29:21,440
university or your ranking class, how that affects 
your wage premium, but as you can already imagine,  

284
00:29:21,440 --> 00:29:25,840
this major and the ranking all have 
a disadvantage, right, as I said,  

285
00:29:25,840 --> 00:29:31,120
you are the worst basically if you are above 
the cut off your worst among the eligible group.  

286
00:29:31,120 --> 00:29:35,520
And if you include them that only makes 
your weight premium slightly larger.  

287
00:29:37,120 --> 00:29:42,160
If you compare, so as I said before now, 
you have, if you have above the cut off,  

288
00:29:42,160 --> 00:29:46,320
you're more eligible, you're more likely 
to go to university go to elite university,  

289
00:29:46,320 --> 00:29:51,600
but you're less likely to go to, to major 
in a popular field. So if you consider that,  

290
00:29:52,160 --> 00:29:57,520
uh take care of that, you'll find actually the 
wage premium or elite education slightly higher,  

291
00:29:57,520 --> 00:30:01,200
from six, to include, to be 
seven percent, for example,  

292
00:30:02,240 --> 00:30:08,880
uh and now come, you know, you, I show you that 
yes elite education now has some wage premium,  

293
00:30:08,880 --> 00:30:15,120
however, does that mean that a young person 
now become elite in this society? If you know  

294
00:30:15,120 --> 00:30:21,040
about this country, you would say, oh I don't 
know, because many elite dimensions are not,  

295
00:30:21,040 --> 00:30:25,920
cannot be captured by the wage premium. 
For example, you know, maybe going to a  

296
00:30:25,920 --> 00:30:32,400
banking industry, finding a job in SOE or 
have a hu kou (户口) in Beijing or Shanghai,  

297
00:30:32,400 --> 00:30:38,160
these are typically also very important for elite 
status. Does going to elite university change all  

298
00:30:38,160 --> 00:30:44,160
these dimensions? That's what we all want to 
answer. Uh so we examine like three dimensions  

299
00:30:44,160 --> 00:30:50,400
to shed light on this question. First, we have to, 
we look at how the job, which job characteristics  

300
00:30:50,400 --> 00:30:56,480
explain the wage premium. Is that to be explained 
by your occupation industry, or ownership? Or is  

301
00:30:56,480 --> 00:31:01,920
actually explained by something like within the 
indiv- within the industry occupation? So you're  

302
00:31:01,920 --> 00:31:09,440
still working in the same occupation, but you have 
different wage. We also define elite occupation  

303
00:31:09,440 --> 00:31:15,280
industry and ownership in an interesting 
way, I think. So in the data we ask people:  

304
00:31:16,400 --> 00:31:23,280
what is their ideal occupation industry? We know 
they real-life industry and occupation, right,  

305
00:31:23,280 --> 00:31:28,640
we know their ideal industry and occupation, then 
this gives us some ratio. If an industry like  

306
00:31:28,640 --> 00:31:36,480
banking, most people hope to work in, but very 
few can realize this job, then this is likely  

307
00:31:36,480 --> 00:31:43,520
to be an elite occupation similarly for industry 
and the ownership. So this gives us some measure  

308
00:31:43,520 --> 00:31:50,000
of eliteness of their job, and finally we also 
check some non-wage benefits like whether they  

309
00:31:50,000 --> 00:31:55,920
get houses subsidy, whether they have like all 
types of insurance. Uh in the end, we don't find,  

310
00:31:56,560 --> 00:32:03,600
or in all this dimension, we don't 
find evidence that this means the the  

311
00:32:03,600 --> 00:32:10,560
people with elite education or also enter the 
elite club. So don't worry about the tables,  

312
00:32:10,560 --> 00:32:16,800
if we do, if we check which job characteristics 
explain the wage premium, in the end we find is  

313
00:32:16,800 --> 00:32:23,200
not explained by their occupation or industry. 
It's actually explained by some variation within  

314
00:32:23,200 --> 00:32:28,800
occupation industry, so you are not necessarily 
more likely to work for the bank industry, or more  

315
00:32:28,800 --> 00:32:34,640
likely to work for the government. It's within the 
same industry, you earn a higher wage. So if we  

316
00:32:34,640 --> 00:32:42,560
define industry or occupation uh by this ratio, 
I mentioned it to you, maybe this, uh this is  

317
00:32:42,560 --> 00:32:49,200
a bit interesting, uh so for example, giving like, 
managed person, management personnel or being a  

318
00:32:49,200 --> 00:32:55,520
leader in a firm or in the government, uh is a 
occupation, you know, you typically think this  

319
00:32:55,520 --> 00:33:02,560
as in like elite occupation, we can also measure 
this with data. For example only one percent of,  

320
00:33:02,560 --> 00:33:08,560
of people would realize a management personnel 
occupation, but if you look at how many people  

321
00:33:08,560 --> 00:33:15,440
hope to work as, in this occupation, it's 10 times 
high. So this ratio is only 0.1, meaning that only  

322
00:33:15,440 --> 00:33:23,040
one in 10 of people could realize their dream. But 
if you look at a skilled worker, or for example,  

323
00:33:23,040 --> 00:33:30,080
uh you'll find a ratio, of if you look at clerks, 
for example you'll find the ratio would be four,  

324
00:33:30,080 --> 00:33:39,840
so you know only, only 7 percent of people would 
like to work as a clerk, but 28 of people have 28%  

325
00:33:39,840 --> 00:33:44,240
had to work as a clerk in the end. 
So this gives us this (unintelligible)

326
00:33:45,360 --> 00:33:52,960
elite status or occupation. Similarly 
we can look at uh industry or ownership,  

327
00:33:52,960 --> 00:33:59,600
and we don't find that this is affected by being 
above the cutoff or not. If we also look at other  

328
00:33:59,600 --> 00:34:05,520
dimensions like whether you have a like hukuo
in Beijing or whether you get housed in subsidy  

329
00:34:05,520 --> 00:34:10,000
or whether you have better insurance, 
all these are not, we find no difference  

330
00:34:10,720 --> 00:34:17,120
between above the cutoff or below the cutoff. So 
all the message is, yes we find the wage premium,  

331
00:34:17,120 --> 00:34:24,320
but if you measure like elite status in other like 
typical sociology career, you wouldn't find that  

332
00:34:24,320 --> 00:34:29,440
being above the cutoff is enough for 
you to become an elite in the society.  

333
00:34:30,960 --> 00:34:36,480
I wouldn't go to the details but, but we don't 
we only have the first job in our survey,  

334
00:34:36,480 --> 00:34:40,480
right? By our design, we cannot 
follow these people in labor market,  

335
00:34:40,480 --> 00:34:44,640
uh and to get some sense to understand 
how important the first job is,  

336
00:34:44,640 --> 00:34:51,840
we downloaded like millions of series are from a 
major job recruitment site called dropping.com,  

337
00:34:52,640 --> 00:34:58,800
and there people have their CV and job history, 
and we look at the correlation of how the first  

338
00:34:58,800 --> 00:35:04,240
job affects their future job wages. And we 
find strong persistence. And this is also true  

339
00:35:04,240 --> 00:35:09,440
in the U.S., I guess. Your first job is very 
important determine your future job earnings.  

340
00:35:11,120 --> 00:35:17,120
So so the message is that even though we 
don't have future jobs for these individuals,  

341
00:35:17,120 --> 00:35:24,720
our findings is likely to matter for the future 
too. So come to the social mobility part, it might  

342
00:35:24,720 --> 00:35:30,400
be sounds a bit confusing if on the one hand yes, 
you know you earn a high wage, of course you have  

343
00:35:30,400 --> 00:35:35,920
a mobility now right because your wage rank is 
increased. But when people talk about social  

344
00:35:35,920 --> 00:35:41,120
mobility, there's another important dimension 
is this intergenerational mobility. As you,  

345
00:35:41,120 --> 00:35:47,120
we all know, there's always your father like your 
parents status and child your status is always  

346
00:35:47,120 --> 00:35:53,360
positively correlated there's always a positive 
line between parental status and income status,  

347
00:35:53,360 --> 00:36:00,240
where we won't earn the children's status. We want 
to know whether this slope changed by being above  

348
00:36:00,240 --> 00:36:08,400
or below the cutoff. This is what I mean. So just 
to give you two scenarios uh now each the red line  

349
00:36:08,400 --> 00:36:15,200
indicates those above the cutoffs and the blue 
line indicates those below the cutoff, and in the  

350
00:36:15,200 --> 00:36:24,320
first scenario, uh the this red line is flatter 
than the blue line, so I forgot to say, the x-axis  

351
00:36:24,320 --> 00:36:31,760
is the income rank of the parent and the y-axis is 
the income rank of the children, In every society  

352
00:36:31,760 --> 00:36:37,680
there's a positive slope. If your parents have a 
high rank, the children tend to have a high rank.  

353
00:36:37,680 --> 00:36:43,440
What's interesting here, we are interested in, if 
you above the cut off, if this rank become flatter  

354
00:36:43,440 --> 00:36:50,080
or steeper, it becomes flatter meaning that now 
the parental income becomes less important if you  

355
00:36:50,080 --> 00:36:56,560
are above the cutoff. This is called increased 
social mobility, intergenerational mobility.  

356
00:36:56,560 --> 00:37:03,440
And in the other case if you are above the 
cutoff, actually this line become even steeper,  

357
00:37:03,440 --> 00:37:09,600
so it's like the parental status becomes even more 
important. I think this is what Pierre Bourdieu  

358
00:37:09,600 --> 00:37:14,880
might talk about, you know, elite parents would 
have more elite kids, you know, around the elite  

359
00:37:14,880 --> 00:37:21,440
education. It's just a way for elite reproduction, 
uh and we need to see that exactly we don't know  

360
00:37:21,440 --> 00:37:26,880
which is likely to be true. It depends on your 
thinking about the society. If you think the  

361
00:37:26,880 --> 00:37:32,960
exam system is great, it levels leverage levels 
the play field, and the parental status should  

362
00:37:32,960 --> 00:37:38,160
become less important once they're above the 
cutoff. If you think oh in the end there's a  

363
00:37:38,960 --> 00:37:45,440
the rich kids should learn even more given the 
same education, you would likely to expect this is  

364
00:37:45,440 --> 00:37:52,400
the second case. And what we find in the data 
actually is neither the case. So that's why it's  

365
00:37:52,400 --> 00:37:58,800
interesting to see the data, uh because you know 
if you if this is so important, such important  

366
00:37:58,800 --> 00:38:03,920
institution in China, if you search online, 
you find scholars have all types of argument  

367
00:38:03,920 --> 00:38:10,080
and hypothesis. Some think or it's very good 
system especially for the poor, and something is  

368
00:38:10,080 --> 00:38:15,200
basically a illusion, it's cheating. In the end, 
it doesn't help the poor. So it's really important  

369
00:38:15,200 --> 00:38:20,960
to show the data and this is what we find. 
You basically see a parallel shift which means  

370
00:38:21,680 --> 00:38:28,400
being above the cutoff does neither increases nor 
decreases inter-generational link. There are few  

371
00:38:28,400 --> 00:38:36,160
messages here. First, you know, if you think 
uh the exam system attenuates parental status  

372
00:38:36,160 --> 00:38:41,600
parental income influence, then this is the 
illusion, this is too naive, uh this parental  

373
00:38:41,600 --> 00:38:47,520
status is still very important. One way to look at 
this is to think about if you are per case here,  

374
00:38:47,520 --> 00:38:56,960
for example, at the the 20 or 40, 20 to 40 percent 
income, 20 to 40 quantile income, and your,  

375
00:38:56,960 --> 00:39:03,600
if you're above the cutoff, right, you indeed have 
a higher wage rank. But if you compare with the  

376
00:39:03,600 --> 00:39:10,560
kid from the very rich, the top 20 percentile 
income families, and even he or she doesn't  

377
00:39:10,560 --> 00:39:17,760
pass the exam, his wage rank still higher. So now 
you so coming back whether can I compete with the  

378
00:39:17,760 --> 00:39:24,640
the second generation rich, now the answer is it 
depends on how rich the the second generation is.  

379
00:39:24,640 --> 00:39:31,840
If he is from the middle group, yes you can you 
can compete with him or her now. But if he is  

380
00:39:31,840 --> 00:39:39,840
very very rich family, you still cannot compete 
with her or him. But I want to be clear, this is  

381
00:39:39,840 --> 00:39:46,960
not too pessimistic. In many elite institutions, 
actually it's likely to be this case. You know,  

382
00:39:46,960 --> 00:39:53,120
if you go to elite university, if your parents 
are richer, you earn even more, right? So this  

383
00:39:53,120 --> 00:39:59,840
is more optimistic than this case, but more 
pessimistic than the level playing field argument.  

384
00:40:02,240 --> 00:40:06,240
Finally we want to understand 
what drives the wage premium.  

385
00:40:07,600 --> 00:40:12,880
Is that driven by human capital? You know, you 
earn more, you're more productive so you get  

386
00:40:12,880 --> 00:40:18,640
paid more. Or is that driven by social networks? 
You have, you know, you know better students,  

387
00:40:19,280 --> 00:40:24,400
father, your classmate's father has worked for 
the government, you get a better job. Or it's  

388
00:40:24,400 --> 00:40:31,360
actually a signaling. You do not necessarily learn 
more, but actually you'll still get a higher wage.  

389
00:40:31,360 --> 00:40:37,120
So this is very difficult to to pin down, so 
what I show you is some suggestive evidence,  

390
00:40:38,080 --> 00:40:41,840
and you know if there are very few studies 
trying to pin down each channel because it's  

391
00:40:41,840 --> 00:40:47,280
very difficult to measure any of them as you 
can imagine. What is human capital, for example?  

392
00:40:48,080 --> 00:40:54,080
so what we do is we which the typical matter of 
human capital is your performance in colleges,  

393
00:40:54,880 --> 00:41:00,880
and you want to ideally measure some tests 
that everyone takes across universities,  

394
00:41:00,880 --> 00:41:07,040
so that is comparable. So gpa is not great right, 
because every different university use different  

395
00:41:07,040 --> 00:41:13,840
ways of gpa diff- like given gpa, so it's not 
very comparable. So there's one measure that is  

396
00:41:13,840 --> 00:41:18,720
close to this idea or most student would pick, 
and they are standardized so you can compare  

397
00:41:19,840 --> 00:41:25,600
across universities. That is the National English 
Test every student, almost every student would  

398
00:41:25,600 --> 00:41:31,520
would take it, and the scores are standardized, 
and it's I think now graded by computer,  

399
00:41:31,520 --> 00:41:38,240
so it's comparable. And then we have different, 
many other measures, in terms of certificate or  

400
00:41:38,240 --> 00:41:45,760
exam performance in college, and we check whether 
there's difference for those above the cutoff and  

401
00:41:45,760 --> 00:41:52,720
below cutoff, and we don't find any difference. 
Also, we ask them how many hours they study in cl-  

402
00:41:52,720 --> 00:41:58,000
in college, and they don't study for for 
a lot, but if they study, they often study  

403
00:41:58,000 --> 00:42:03,200
English. So we also look at how they put their, 
you know, how their, how their efforts look like  

404
00:42:03,200 --> 00:42:08,640
for those above the cutoff and below the cutoff. 
And we find no evi- like no difference either.  

405
00:42:08,640 --> 00:42:14,240
So in all the measures of human capital, like 
majors, we cannot explain our finding either.  

406
00:42:14,240 --> 00:42:18,960
So in all these measures, we can measure, 
or we can have in terms of human capital,  

407
00:42:18,960 --> 00:42:25,280
we find no difference. So it's not that we want 
to reject this hypothesis, but we just have no  

408
00:42:25,280 --> 00:42:32,640
evidence that we perform better, or have more uh 
like more certificate to prove they're better.  

409
00:42:33,920 --> 00:42:38,560
And if we look at network, you know, if you think 
if you know a little bit China, about China,  

410
00:42:38,560 --> 00:42:45,200
you would think about this is about network, uh 
so what we do is, we look at how, we ask them  

411
00:42:45,200 --> 00:42:51,200
how you find your job, we also look at how your 
networks affect their wage. So when we come to  

412
00:42:51,200 --> 00:42:58,640
the job search part, people usually use different 
ways of search jobs like off-campus job fairs  

413
00:42:58,640 --> 00:43:04,800
or on-campus job fairs or job website or 
using networks and we don't find anything  

414
00:43:04,800 --> 00:43:10,880
that those above the cut-off are more likely to 
use evidence, uh use network. If anything, we  

415
00:43:10,880 --> 00:43:17,120
find they're more likely to use on-campus fairs, 
but that's a bit related to what Pang Le said.  

416
00:43:17,120 --> 00:43:22,880
Those elite universities also attract better 
employers, so they have on-campus fairs which  

417
00:43:22,880 --> 00:43:28,640
gives these people more opportunities to get 
good, better jobs, but this is not a network  

418
00:43:28,640 --> 00:43:33,600
channel, this is more likely to be a signaling 
channel in our understanding if the reputation  

419
00:43:33,600 --> 00:43:40,960
of the university matters in terms of attracting 
the employers. And if you look at the, how this  

420
00:43:40,960 --> 00:43:47,360
network affect, uh wage premium, so what, so 
to be sure, yes if you are above the cut-off,  

421
00:43:47,360 --> 00:43:52,960
you go to elite university, indeed you have better 
networks. For example, if you measure the share of  

422
00:43:52,960 --> 00:43:59,200
your classmate whose father has college degree, or 
whose father has a, is a party member, indeed the  

423
00:43:59,200 --> 00:44:04,160
probability is higher. So you know, if you're 
above the cut off, your classmate background  

424
00:44:04,160 --> 00:44:09,920
is better, that's for sure. We do find that in the 
data, but we found this cannot explain your wage  

425
00:44:09,920 --> 00:44:17,280
premium, so it's not explained by your part, your 
your government background. But there's one caveat  

426
00:44:17,280 --> 00:44:23,600
in these findings. We're looking at the first job, 
and this type of weak ties, you know, your your  

427
00:44:23,600 --> 00:44:29,440
classmates, your schoolmate's father's connection 
often matters in the long run. In the first job,  

428
00:44:29,440 --> 00:44:35,120
maybe it doesn't matter much, but in the long run, 
it can meant more, that well we cannot study that.  

429
00:44:36,160 --> 00:44:41,520
And another, another piece of evidence on 
signaling is suggestive, because it's very  

430
00:44:41,520 --> 00:44:46,400
difficult to measure signaling unless you 
are in a lab, you know. In the real world,  

431
00:44:46,400 --> 00:44:52,480
you don't really know how people send the signals. 
We ask them something about job discrimination,  

432
00:44:52,480 --> 00:44:57,200
what kind of discrimination they experience 
uh when they are on the job market,  

433
00:44:57,200 --> 00:45:04,080
like we ask like gender, uh appearance, 
(unintelligible), rural, etc. We don't find any any  

434
00:45:04,080 --> 00:45:10,800
difference in this dimension. There's only one uh 
dimension which is very sizeably different but not  

435
00:45:10,800 --> 00:45:16,400
precisely estimated, is that if you're above 
the cutoff you're less likely to see that  

436
00:45:16,400 --> 00:45:21,760
your, you are discriminated because of a 
university rank. This is almost mechanical, right,  

437
00:45:21,760 --> 00:45:26,480
because if you're above the culture of you are 
more likely to go to elite university and you're  

438
00:45:26,480 --> 00:45:32,480
less likely to think you'll be discriminated, or 
you are discriminated because of the university  

439
00:45:32,480 --> 00:45:39,360
rank. And, but this is all, this you know, if you 
think discrimination is a negative signaling,  

440
00:45:39,360 --> 00:45:44,960
then this is also more consistent with the 
signaling mechanism. So in this part, it's not  

441
00:45:44,960 --> 00:45:50,480
that we want to reject any hypothesis, we don't 
need and this could be a paper without this part,  

442
00:45:50,480 --> 00:45:57,680
what we want to just to know, we just, let's have 
try as many measures as possible and see which  

443
00:45:57,680 --> 00:46:03,440
is more consistent with the data. And in the end 
it's more consistent with the signaling mechanism,  

444
00:46:03,440 --> 00:46:09,840
and in reflection, in retrospect I wasn't very 
surprised. Initially I was a bit disappointed,  

445
00:46:09,840 --> 00:46:15,520
oh we don't find human capital in this, but now 
I kind of think, you know, you think about this  

446
00:46:15,520 --> 00:46:22,720
education institution, which is very strict 
in interest, but very relaxed exit,  

447
00:46:22,720 --> 00:46:27,920
so you know students don't have a lot 
of incentive to learn much. I don't  

448
00:46:27,920 --> 00:46:32,160
know whether you have studied in Chinese 
college whether you shared my feelings,  

449
00:46:32,160 --> 00:46:37,360
but they did they don't have much incentives 
to to to learn for graduation, for example.  

450
00:46:38,160 --> 00:46:46,880
Uh so yeah I can, I can stop here but 
feel free to ask questions. Thanks.