WEBVTT

1
00:00:00.001 --> 00:00:04.840
good morning to this very intimate session about the innovations of

2
00:00:05.110 --> 00:00:10.930
so gender research, I will be presenting my work about

3
00:00:11.970 --> 00:00:20.280
intersectionality, resources and benefit received. We've changed the title slightly, but it's only on the front end, in which we will be

4
00:00:20.610 --> 00:00:25.479
studying the role of resource and shaping intersection inequalities in benefit receipts.

5
00:00:27.260 --> 00:00:35.890
so prior research has already established that there are inequalities along the lines of gender H. A. Migration background to benefit receipt.

6
00:00:35.940 --> 00:00:47.779
And in our own prior research we've also found that at the intersection of gender, migration, background, and age. Some groups are disproportionately more vulnerable for benefit received.

7
00:00:48.719 --> 00:00:49.850
and that

8
00:00:50.030 --> 00:00:55.040
and through processes of conversation.

9
00:00:55.380 --> 00:01:01.230
amplification, and attenuation. These inequalities can basically be non- additive.

10
00:01:01.960 --> 00:01:03.750
The question ever remains.

11
00:01:03.830 --> 00:01:07.970
water is at play here. So in this paper we will be studying

12
00:01:07.989 --> 00:01:11.040
the mechanism that might allow these inequalities.

13
00:01:11.240 --> 00:01:21.250
 and in this paper we will be swimming in on resources. So it might be that these inequalities for these intersectional groups are caused by.

14
00:01:21.420 --> 00:01:27.950
and inequalities and resources, such as economic capital, social capital, cultural capital and person capital.

15
00:01:29.270 --> 00:01:30.270
And

16
00:01:30.410 --> 00:01:45.550
the aim of this paper is basically 2 faults. First, we will establish that. And resources are basically mediating benefits received along the lines of gender migration by ground and H, and the second aim is to test whether or not these mechanisms are also intersectional.

17
00:01:47.810 --> 00:01:53.530
So our theoretical framework basically sums up into this picture.

18
00:01:53.630 --> 00:01:58.059
Let me just discuss how these capital forms might actually work.

19
00:01:58.120 --> 00:02:02.069
Maybe economic capital

20
00:02:02.360 --> 00:02:04.020
denote what you have.

21
00:02:04.310 --> 00:02:11.760
these are resources that are directly monetizable, such as your wealth or income, but also forms of human capital like education.

22
00:02:13.030 --> 00:02:17.710
And then cultural capital, basically, and who you are.

23
00:02:18.270 --> 00:02:31.710
And this is basically in a budget Asian tradition, looking at the ways people signal their status, but also through language, and how proficient they are in this case. Dutch language.

24
00:02:31.940 --> 00:02:52.790
then, social capital, who do you know? And social capital basically kept us all the resources that are in your network. this can either be information for for for for job opportunities. your friends can help you with finding jobs, but also people can vouch for you and and and over helping you to find forms of employment

25
00:02:52.970 --> 00:03:00.469
and person capital. And how you look. And this is basically stepping into the

26
00:03:01.210 --> 00:03:08.139
barito's work about how physical traits are very important on the labor markets and helping you to find jobs as well.

27
00:03:08.350 --> 00:03:13.199
In this case we will be mainly looking at. Forms of health such as general health and mental health.

28
00:03:13.270 --> 00:03:16.860
People that are more healthy are

29
00:03:17.050 --> 00:03:22.529
are generally more productive, and therefore might be more suitable top candidates for an employer.

30
00:03:25.060 --> 00:03:41.489
and I don't have the time to discuss all mitivocal arguments here. My my hypotheses are presented here in this table. And I'd like to pick this one out here, which is basically the first generation migrants.

31
00:03:41.900 --> 00:03:58.180
They generally have the least forms of wealth and income, and generally these migrants enter the Netherlands to find employment here, and they have generally lower status jobs having a lower income, and therefore are less likely to generate the wealth over the live course.

32
00:03:58.570 --> 00:04:03.730
 And the second case.

33
00:04:03.900 --> 00:04:22.750
and for the second-generation migrants They, although they have grown up in the Netherlands, they generally still have less resource when that works, and therefore might be less likely to find stable employment, and therefore might be more likely to end up in

34
00:04:22.920 --> 00:04:26.140
some form of benefits and over your life course

35
00:04:28.770 --> 00:04:37.979
than in the sectionality. And the quality of of intersectionality is that multiple sources of disadvantage and also advantage are mutually and influential.

36
00:04:38.280 --> 00:04:49.139
and therefore that they should not be studied in addition or in isolation to each other, but should always be considered in relation to each other, because

37
00:04:50.280 --> 00:04:55.069
combining 2 forms of disadvantage, for instance, can lead to an increased level of

38
00:04:55.140 --> 00:04:56.860
and disadvantage.

39
00:04:57.420 --> 00:05:01.370
and also therefore leading to higher levels of benefit received.

40
00:05:02.100 --> 00:05:03.770
and

41
00:05:04.520 --> 00:05:10.489
therefore, like relating it back to already the status quo of research in the

42
00:05:10.820 --> 00:05:22.609
social welfare research, it basically adds an additional component of either an advantage, or an or decreased advantage compared to what an additive model would

43
00:05:22.940 --> 00:05:23.990
would bring you.

44
00:05:26.290 --> 00:05:32.779
So these are all my intersectional hypotheses, and I also, for the sake of time will not be discussing every one of them. But just

45
00:05:32.880 --> 00:05:42.950
to highlight our theoretical argumentation for second generation, my good women, we find that they actually outperform on

46
00:05:43.140 --> 00:05:44.090
in

47
00:05:44.300 --> 00:05:54.989
education, and generally have a higher educational attainment. So we already know that women outperform men in education, but it might be especially the case. So for second-generation women.

48
00:05:55.110 --> 00:05:56.100
and

49
00:05:56.770 --> 00:06:04.779
compared to second-generation man, the gap between the educational payments

50
00:06:04.910 --> 00:06:09.340
between men and women, 4 migrants might be bigger than it is for Dutch natives.

51
00:06:09.360 --> 00:06:14.890
and therefore this might actually help them to be less vulnerable for a benefit receipt.

52
00:06:17.020 --> 00:06:31.179
We have been studying this using a combination of 2 data sources. We have access to longitudinal register data through the center of your statistics in the Netherlands. And we've combined it with this household panel data.

53
00:06:31.290 --> 00:06:45.289
We have sourced information from the person register, from the wealth register, from the income register, but also from a register which denotes whether or not. They have been receiving a benefit at some point in time.

54
00:06:45.450 --> 00:06:48.630
 then, to

55
00:06:48.720 --> 00:07:12.019
to combine this this data set. Of course, there should have been a link between the registers and the list panel data, and some of the panel data participants have not agreed for this link to be made. So they obviously are not in the data set, but also for 10% of the cases, it was not possible to make a reliable link, and therefore there's also no link

56
00:07:12.020 --> 00:07:20.999
linkage possible. Then we only selected persons between 20 and 60 years old as they formed a core workforce in the Netherlands.

57
00:07:21.380 --> 00:07:28.769
then we should at least have had 2 observations, so that we can run longitudinal models and look at changes over time.

58
00:07:29.120 --> 00:07:40.449
and then we only selected one person per household. As I said the list panel was a household survey, and to prevent an overestimation of the effect, we selected just one member of the household.

59
00:07:41.350 --> 00:07:44.249
In the end we arrived at the

60
00:07:44.410 --> 00:07:51.700
an analytical sample of about 4,000 individuals, with about 24,000 observations.

61
00:07:51.920 --> 00:07:57.269
ranging from 2,008 up until 2,019,

62
00:07:57.300 --> 00:08:07.789
and we excluded the years of 2,020 and 2,021, and we did not want to have any influences of the COVID-19 pandemic to play a role in our analysis.

63
00:08:08.910 --> 00:08:14.909
So let's look at the operationalizations. Of of course you will be looking at benefits received which we have

64
00:08:15.260 --> 00:08:26.109
rationalized as whether someone has at least received either social assistance or unemployment insurance for at least one month per calendar year.

65
00:08:26.660 --> 00:08:39.739
Then the demographic variables and gender was sourced from the person register in 2,019 and age groups. And we have made, based on the age of the first of January in 2,014

66
00:08:39.950 --> 00:08:54.579
and we have identified 4. So people in the twenties, thirties, forties, and fifties and we are categorizing them. because that's needed for the creation of these intersectional groups that people will be looking at.

67
00:08:55.330 --> 00:08:57.100
and the migration background

68
00:08:57.180 --> 00:09:10.300
for which we are both looking at origin. So, people having either a Dutch, Western, or non-western, I see they're still a little bit a little bit of Dutch in my slides and on Western

69
00:09:11.280 --> 00:09:19.860
origins, and then also, whether or not the people from the people that have a migration background, or the first generation, or descendants from people that migrated to the Midlands.

70
00:09:21.470 --> 00:09:39.520
and then we consider a large array of resources that people may have, and so for wealth and income, which was constructed as a formative latent structure, and which combines basically and individuals. Income, wealth and home ownership in each year.

71
00:09:39.800 --> 00:09:52.840
and education was also personalized as education in years which then announced the number of years in which an individual should have at least spent an education to reach that level of education

72
00:09:53.960 --> 00:10:04.140
and cultural capital measures, highbrow cultural activities, and also the number of times that people attended such activities. And again, this was also

73
00:10:04.620 --> 00:10:17.540
formed using a formative latent structure. The language proficiency basically indicates whether or not in person, has trouble reading, writing, and speaking, and Dutch social capital, and

74
00:10:17.890 --> 00:10:36.600
combines 3 aspects of social capital. using also again a formative latent structure which combines basically the number of highly educated people in the network, and the people of the the number of people that are full time employed. The size of the network, but also the density of the network.

75
00:10:37.340 --> 00:10:40.710
And this is the core network of individuals. So first.

76
00:10:40.800 --> 00:10:43.710
the respondents were asked to list, fight

77
00:10:43.800 --> 00:10:52.709
5 people in the core network, and then subsequently, they were asked whether or not this conduct was then highly educated or and full time employed.

78
00:10:53.550 --> 00:11:13.119
and general health was, again, combined using a formative latent structure. combining the self reported general health, the Bmi and also a retrospective question in which they indicated whether or not they were unable to work because of health reasons.

79
00:11:13.450 --> 00:11:22.570
The last question was made you using a reflective, latent, and latent structure, and it combines 5 questions about mental health.

80
00:11:25.470 --> 00:11:32.350
To test our policies. We used Bayesian not to level structure creation models. And we.

81
00:11:32.550 --> 00:11:48.120
this modeling approach is basically inspired by what is considered to be in the state of the art analysis in quantitative intersectional research. Mahida models, and which is the multi-level analysis of individual heterogeneity and discriminatory accuracy.

82
00:11:48.250 --> 00:11:53.889
And we take the core idea of this methodology of

83
00:11:54.060 --> 00:11:58.319
calculating both an additive effect and a non- additive effect.

84
00:11:58.390 --> 00:12:14.620
instead of doing it in one model, and we are now doing it with 2 separate models. We calculate an additive model and an intersectional model, and then to calculate non-editive effects. We subtract the predictions of the editative model, and from the predictions of the intersectional model.

85
00:12:14.860 --> 00:12:22.049
and we do that to be able to test the mediation effects, and then also to have

86
00:12:22.070 --> 00:12:26.700
a norm additive, and an additive component for this mediation effect in our models.

87
00:12:28.520 --> 00:12:37.709
As you can imagine, these models can be quite complex. So just to illustrate, this is basically the full model

88
00:12:37.830 --> 00:12:39.430
calculating the

89
00:12:39.440 --> 00:12:44.930
in the sectional mediating effects leading to benefit received.

90
00:12:46.260 --> 00:13:08.599
So let's look at the results. First, I will be looking at whether or not there are intersectional patterns in benefit to. So on the last panel you see the incidences of benefits received in the Netherlands for all the intersectional groups. And these intersectional groups are basically all the unique combinations of the demographic variables that I have just discussed.

91
00:13:09.020 --> 00:13:12.980
 and there's a large variation in

92
00:13:13.160 --> 00:13:20.740
benefit deceit. The lowest group has not 0 point 0 7% incidence of benefit receipt.

93
00:13:20.960 --> 00:13:24.160
which I'm looking at my notes here.

94
00:13:28.410 --> 00:13:39.349
which are the second generation, non-western women in their 40 s. And the highest incident is among the first-generation and Western women in their 20 s.

95
00:13:40.880 --> 00:13:56.960
And after we have been calculating the intersectional effects. So we have run a model, including all the edit. The additive effects of gender immigration background and H, and we so, and we subtracted that from an intersectional model

96
00:13:57.190 --> 00:14:07.749
for which we found only one group to have an intersectional effect in this case, being the first generation of women, and with a Western origin in their 20 s.

97
00:14:09.270 --> 00:14:12.840
And this is in contrast to our

98
00:14:12.990 --> 00:14:30.400
or earlier findings, and it is, I think, somewhat to do with the limited sample size that we have, or our initial paper was on about 200,000 individuals, and due to to the linkage and of to the list panel data, and the apparently I in just 4,000 individuals here.

99
00:14:32.160 --> 00:14:42.319
then, before we will be looking at the intersection of the intersection. Mediating effects.

100
00:14:42.620 --> 00:14:46.070
 intersectional main

101
00:14:46.240 --> 00:15:02.069
to the mediating effects. Let's just first look at the main effects of these resources on benefits. And what we do see here in this thought is that wealth is largely determining to. I don't know if people are receiving a benefit in the Netherlands.

102
00:15:02.210 --> 00:15:03.280
and

103
00:15:03.680 --> 00:15:24.760
where, on average, in our sample, and people receive so 7.2 and right? So but 2% receives and benefits per year, we do find that with each increase of wealth and income, people actually have a reduction of about 9. So that's quite a large effect

104
00:15:24.940 --> 00:15:30.430
for education, we do find an effect when we are not controlling for any other

105
00:15:30.780 --> 00:15:46.149
resources. So the blue points denoted prediction, using a model where only the resource is modeled individually as a mediating factor. And the orange dots are basically models that control for all the

106
00:15:46.250 --> 00:15:48.599
for all the resource effects simultaneously.

107
00:15:49.170 --> 00:15:55.480
We do not find this education effect. When we have controlled for all other resources.

108
00:15:55.530 --> 00:15:57.470
which

109
00:15:57.670 --> 00:15:59.479
indicates the

110
00:15:59.530 --> 00:16:08.760
the mutual correspondence of these resources, it might be that the education effect mainly operates, for instance, through health or through

111
00:16:08.790 --> 00:16:18.169
cultural capital, for which we find effects while controlling for all the other effects of of resources.

112
00:16:18.910 --> 00:16:31.140
So, looking at the bigger picture. When controlling for all the resources, we find that mental health, cultural capital and wealth and income are influential factors.

113
00:16:31.150 --> 00:16:32.360
Oh.

114
00:16:33.410 --> 00:16:50.419
I'm already a bit beyond time. I can have one or 2 more minutes. Okay? And so let's look at the indirect effects and for wealth and income, we find, and basically no gender differences, but we do find differences along the lines of migration background and H,

115
00:16:52.000 --> 00:17:19.290
and for cultural capital we find few deviations. So again, there are no gender differences. And but there are differences between and the first generation groups. people with the first innovation, non-western backgrounds have more cultural capital, and therefore, a lower amount of benefit received, and people with the best immigration background generally have less cultural capital leading to higher incidences, but incidents as a benefit received.

116
00:17:19.400 --> 00:17:24.749
and also for the only the eldest H group, we find an indirect effect

117
00:17:25.820 --> 00:17:47.060
for mental health. We we do find gender differences. So women, on average report to poor mental health, and that might lead to a high incidents of benefit. Receipts in the migrant groups have also poor mental health, increasing their likelihood for benefit receipts and all their age groups have a better mental health leading to lower levels of benefic disease.

118
00:17:48.200 --> 00:17:51.590
So to some out we can

119
00:17:51.770 --> 00:18:00.730
explain gender differences through differences in personal capital, migration backgrounds and age and through differences in economic cultural capital and personal capital.

120
00:18:01.030 --> 00:18:03.919
Then there are just very few intersectional results.

121
00:18:04.160 --> 00:18:27.760
and I will just highlight one case for the sake of time. for instance, and non-western man with the second generation migration background in their fiftys, and we find them to have more cultural capital than we would have expected on additive grounds. So they have disproportionately a lot of cultural capital, and then reducing their benefit received.

122
00:18:29.070 --> 00:18:40.010
So there are 3 main conclusions which I will be skipping over. So wealth, income, culture, capital, and mental health partially explain differences between age, migration, background, and gender.

123
00:18:40.220 --> 00:18:51.400
and we do find some intersectional effects as they have disproportionate disparities in resources, amplifying or reducing their incidents or benefit receipts.

124
00:18:51.570 --> 00:18:55.410
And but yet these non additive indirect effects are few and also very small.

125
00:18:56.510 --> 00:18:59.390
and this has direct implication for social policy.

126
00:18:59.430 --> 00:19:10.910
and, namely, it identifies mechanisms of in the sexual inequalities, in benefit to seed which could inform social policy to make targeted interventions, to address one of the root courses of these inequalities.

127
00:19:11.450 --> 00:19:13.409
and thank you so much for your attention.

128
00:19:23.220 --> 00:19:24.830
And

129
00:19:25.340 --> 00:19:27.520
he planned work in the next bigger

130
00:19:27.550 --> 00:19:34.080
to record it. But I kind of like to. Yeah. Mostly. I?

131
00:19:36.140 --> 00:19:38.130
Then showing

132
00:19:40.100 --> 00:19:44.010
layers the next speaker.

133
00:19:44.290 --> 00:19:46.140
you can have.

134
00:19:47.310 --> 00:19:54.829
Okay. Sorry. Sorry. I thought to shout myself, but it's better to to him.

135
00:19:54.950 --> 00:20:23.460
Okay, the the next paper you're gonna be by the G. Cow. The the title of the paper. Occupational gender segregation intra occupational inequality in gender wage gap dynamics in the United States a causal mediation approach. I give you 2020 min because I like to be equal. But anyway, so if you can finish another 50 min. That will be fine.

136
00:20:23.740 --> 00:20:25.000
Floyd's yours.

137
00:20:25.910 --> 00:20:26.819
the the

138
00:20:36.050 --> 00:20:57.139
okay. Good morning, everyone. It's a great pleasure to be here. My name is Gitao, and I come from the Chinese University. I'm a Phd. Candidate from Chinese University of Hong Kong, Department of Sociology, and my my co-author is Professor Julian and my department. She is also one of my advisors.

139
00:20:57.140 --> 00:21:07.250
and thank for the introduction. by Professor Shilahasse. And actually, I changed my title a little bit.

140
00:21:07.430 --> 00:21:11.399
Okay? So the title today is

141
00:21:11.610 --> 00:21:24.900
evaluating occupation-related forces in the dynamics of the gender wage gap in the United States. I remove the term causal as I will mention in the discussion part.

142
00:21:25.620 --> 00:21:26.910
Okay? So

143
00:21:27.100 --> 00:21:29.740
here's a little bit of background.

144
00:21:30.060 --> 00:21:45.459
well, across all Oecd countries. We know that there has been a dramatic convergence in gender roles of men and women

145
00:21:45.660 --> 00:21:50.259
have past centuries as witnessed the closing gender gap

146
00:21:50.340 --> 00:22:09.630
and among across various domains like educational attainment. And nowadays it's a reversal gender gap in educational attainment and also closing gender gaps in labor force, participation in income, wage earnings and in occupation distribution.

147
00:22:10.040 --> 00:22:11.200
So

148
00:22:11.480 --> 00:22:29.680
in general, this is referred to a gender revolution by social scientists. However, we notice all Social scientists notice that this general revolution seems to have start out in the United States after the 1990. S,

149
00:22:30.190 --> 00:22:40.329
okay, so seeing, since our focus today is earnings, let's zoom in for the trend of gender wage gap.

150
00:22:40.560 --> 00:22:54.530
So we calculated this figure based on a very single measure. That is the ratio between medium male wages and the medium. Female wage!

151
00:22:55.620 --> 00:23:07.720
And this is just a raw ratio for for interpretive. for for for simplicity. and you can see that the

152
00:23:07.870 --> 00:23:12.300
gender wage gap started to close since 1,900 and

153
00:23:12.380 --> 00:23:26.769
80 s. And at first it converged very quickly, however. Ever since me. 1995, it gradually so slowed, or even so start.

154
00:23:29.640 --> 00:23:30.950
Okay? So

155
00:23:31.060 --> 00:23:38.339
actually both the rapid convergence and the slowing or stalling out have

156
00:23:38.580 --> 00:23:54.179
attracted much research. Attention. Scholars have paid close attention to the effect of various occupation-related forces. and there are many influential studies, such as

157
00:23:54.260 --> 00:24:12.969
some One explanation is that for the rapid convergence it's mainly it should be mainly attributed to Manx and women's wage, the changing, their wage distribution, and they use the term polarization. And also there's also a

158
00:24:13.190 --> 00:24:22.340
explanation for the slowing that the rising return to overwork in certain occupations.

159
00:24:22.360 --> 00:24:28.820
slow, the gender wage gap, convergence in recent decades.

160
00:24:28.840 --> 00:24:32.529
and many other very

161
00:24:32.660 --> 00:24:39.899
influential explanations. So, however. however, if we.

162
00:24:40.530 --> 00:24:50.939
if we care about the specific roles, the red, the the all the rows played by the occupation-related forces.

163
00:24:51.310 --> 00:25:01.120
During this whole process we have mixed findings due to different operationalizations. I can give 2 examples, first.

164
00:25:01.610 --> 00:25:06.039
like for the changing wage structure

165
00:25:07.220 --> 00:25:15.720
in some study that they found the change in which structure helps to close the gender wage gap over time.

166
00:25:16.120 --> 00:25:35.329
however, some other studies due to different operationalization, they found it expands the genuine gap over time, and also also in. In the the general discourse seems to have an impression that the reduction in occupational gender segregation is the most

167
00:25:35.800 --> 00:25:39.939
contributor for the reduction in Gender Witch Gap.

168
00:25:40.450 --> 00:25:47.989
Okay, so all the previous studies offer insights. But we it's still unclear that

169
00:25:48.000 --> 00:26:01.230
what's the relative importance of those occupation-related factors in explaining the gender wage gap trend. So in other words, it's unclear which

170
00:26:01.250 --> 00:26:09.579
occupation-related forces exerts greater influence on the gender wage gap trend. And to what extent?

171
00:26:12.320 --> 00:26:13.220
Okay.

172
00:26:13.330 --> 00:26:32.309
so I I will first use a heuristic diagram. To illustrate, what do I mean by all the roles played by occupation in the trend of genuine gap? So this is just a heuristic one. Apparently we start from gender. And then, since

173
00:26:32.810 --> 00:26:42.300
different gender, with workers with different gender are allocated disproportionately to different occupations.

174
00:26:42.330 --> 00:26:51.049
This is the first step gender difference, then different occupations has different way to return by occupation.

175
00:26:51.060 --> 00:26:55.310
So collectively they. This is a

176
00:26:55.520 --> 00:26:58.410
this forms 2 path of

177
00:26:59.090 --> 00:27:05.559
that contribute to general wage gap. and also

178
00:27:05.750 --> 00:27:12.000
and also within each occupation, there is a gender wage gap

179
00:27:12.390 --> 00:27:17.979
due to many factors like

180
00:27:18.450 --> 00:27:19.110
Okay.

181
00:27:20.820 --> 00:27:30.180
like different social capital, like the the the loan work, hours, etc.

182
00:27:31.350 --> 00:27:46.589
So this is the heuristic diagram for the static general wage gap. And over time all the factors or the dots in this diagram can change which will contribute to the change in the overall gender wage gap.

183
00:27:47.500 --> 00:27:48.360
Okay?

184
00:27:48.630 --> 00:27:59.940
And also if we zoom in with the role of occupation, we want to have a clean effect. That's

185
00:28:00.090 --> 00:28:11.209
important factors that we already know. We want to also consider the intersectionality like human capital and demographics, like

186
00:28:11.260 --> 00:28:15.420
family status, parental status.

187
00:28:16.320 --> 00:28:22.159
so we can, conditioning on all those factors which will

188
00:28:23.050 --> 00:28:28.139
give us a relatively clean example, a clean effect.

189
00:28:30.080 --> 00:28:33.470
So I just illustrate the goal.

190
00:28:33.540 --> 00:28:59.200
the goal of our inquiry. So, however, we cannot just fulfill this inquiry from the existing literature because of this obstacle. That is, the methodological tools in existing study cannot simultaneously estimate our occupation-related forces. So the current situation is like the left side, the pi in the left side, that

191
00:28:59.510 --> 00:29:09.330
they are overlapping forces. and what we want to do is to separate them and have a clear-cut evaluation of each part.

192
00:29:10.980 --> 00:29:11.740
Okay?

193
00:29:11.910 --> 00:29:28.019
So our solution is to do a extension. based on the existing decomposition methods. So I will not go through all the mathematical details. But the setup is very simple, and

194
00:29:28.690 --> 00:29:43.389
so here is that The the first step we we begin at the assumption that the potential wage for male and female is just the additive product of

195
00:29:43.880 --> 00:30:01.339
which occupation you are in and given your gender and your occupational and your return within this occupation giving your gender. Okay? So from this we can define the

196
00:30:02.740 --> 00:30:07.159
we can define this, a static gender wage gap

197
00:30:07.280 --> 00:30:11.189
at a very simple and intuitive sense.

198
00:30:13.410 --> 00:30:25.500
and after the rearrangement of terms, this is the final The result of our decomposition, and what we want to estimate in the end.

199
00:30:25.630 --> 00:30:33.249
So the left-hand side is the change in the gender wage gap over time, and

200
00:30:33.360 --> 00:30:50.419
on the right side. You can see 5 parts, and they are in the same structure that each component is a over time change weighted by something, and they

201
00:30:50.760 --> 00:30:59.780
they? They are supposed to be very intuitive. For example, the first component is just, you can see that it's

202
00:31:00.080 --> 00:31:01.910
given occupation

203
00:31:02.690 --> 00:31:12.829
at period T. 2, the gap in wage potential wage between men and women, and this

204
00:31:12.840 --> 00:31:23.409
this within occupation, general wage gap. How does it change over time? We represent this by the difference between 2 time periods.

205
00:31:23.430 --> 00:31:33.130
and it's weighted by the composition of occupation, which is the probability of

206
00:31:33.320 --> 00:31:43.290
people getting into this occupation. And for the second one the second one. The overtime change part is just the

207
00:31:44.440 --> 00:31:57.549
properties, the probability of getting into the probability for men get getting to this occupation minus the probability of women getting to this

208
00:31:57.740 --> 00:32:10.790
occupation at time, a recent time, that in the previous time period, and actually this part in this component

209
00:32:11.010 --> 00:32:28.209
is essentially equivalent to the dissimilarity index. So it's a close map between the occupational gender segregation and the how it translated into the contribution to gender wage gap.

210
00:32:28.930 --> 00:32:54.429
And the second term is that, how? How does the change in overall? the return structure, payoff structure of different occupations translated into a gender wage gap. And the fourth term is just that. How does over time some occupation shrink and some occupation expanded? How does this change translate it into a

211
00:32:54.480 --> 00:32:56.439
generate gap.

212
00:32:56.950 --> 00:33:07.160
and the fifth term is the residual part which reflect the self's change in the self-section of between men and women.

213
00:33:08.600 --> 00:33:15.230
Okay. So the feature of our decomposition framework is that compared with

214
00:33:15.490 --> 00:33:16.430
Cob

215
00:33:16.910 --> 00:33:22.999
like a blender decomposition, we has a minimal model assumption

216
00:33:23.050 --> 00:33:39.470
because we didn't start from the OS estimates. We just start from a very basic probabilistic assumptions. And what's more important is that we

217
00:33:39.980 --> 00:33:53.870
divide the role of occupation into mutually exclusive and exhaustive quantities. It's easier for theoretical mapping than the previous methods for our purpose.

218
00:33:57.060 --> 00:34:07.380
The data I use is the Ipum's current population survey. It's well known for the high quality of labor market information.

219
00:34:07.400 --> 00:34:22.629
And it's an annual survey asking retrospective questions. Our analytic sample. We just follow the convention in this field to use the full time 4 year. Wage workers, age 25 to 54.

220
00:34:22.640 --> 00:34:40.600
to exclude the possibility of idiosyncratic selection into retirement. and we have 4 sets of variables. The first is hourly wage, and the second is the to treatment, gender and occupation.

221
00:34:40.800 --> 00:34:56.210
And we also control for human capital qualifications which include the education and the potential experience. And finally, we also control for a a set of or relevant

222
00:34:56.470 --> 00:35:05.420
demographic characteristics, including an age resid location of residents

223
00:35:05.500 --> 00:35:06.970
and

224
00:35:06.979 --> 00:35:14.180
family family status and parental status, meritose status and parental status, etc.

225
00:35:16.740 --> 00:35:17.830
So

226
00:35:18.050 --> 00:35:38.570
here's to remind you that our quantity of interest, what we want to estimate is the I use the shorthand as the delta total, which is the change in the total gender, wage, gap and delta within the changing within occupational gender, wage, gap and

227
00:35:38.609 --> 00:35:49.320
delta sack means segregation! How does changing segregation contributed to a changing genetic gap and delta price means the change in the

228
00:35:49.390 --> 00:35:51.849
occupational wage structure.

229
00:35:51.960 --> 00:35:57.889
and delta comp means the change in the occupational composition.

230
00:35:58.350 --> 00:36:02.759
and finally, delta covariance, which is the residual part.

231
00:36:03.350 --> 00:36:08.669
Okay? So we, as both do the very, very.

232
00:36:09.100 --> 00:36:11.129
This simple descriptive

233
00:36:11.420 --> 00:36:21.130
estimate for estimation, for our majestic gap, which is just everything, is the raw gap we didn't control for

234
00:36:21.180 --> 00:36:32.609
human capital and demographics. Then we also do it for a adjusted gap which controlled for all the possible covariance we can control.

235
00:36:34.180 --> 00:36:35.040
Okay?

236
00:36:35.260 --> 00:36:40.939
And for for the adjusted gap, we estimate them by but

237
00:36:41.350 --> 00:36:53.279
device, machine learning approach because the framework we use that the each component for each component, it's it's

238
00:36:53.580 --> 00:36:59.180
likely it's like to to do a prediction task at the first step.

239
00:36:59.190 --> 00:37:08.250
then plug in to constitute the final estimates. So for the prediction task, it's

240
00:37:09.590 --> 00:37:10.390
okay.

241
00:37:11.770 --> 00:37:27.719
using a machine learning approach is more suitable. And we adopt a D bias machine learning framework to adjust for the second other biases derived from

242
00:37:27.770 --> 00:37:28.950
machine learning.

243
00:37:29.640 --> 00:37:43.270
Okay, so here's the key in ingredients. We use the random forest as our machine learning algorithm. And we also follow the recommended practices.

244
00:37:43.430 --> 00:37:47.160
 out 2, 2 more minutes

245
00:37:48.300 --> 00:37:49.240
through it. Okay?

246
00:37:50.620 --> 00:38:03.630
by Turnojukov, that we do cross-fitting, using 5 folds, and we, after we get the first other. Prediction of all those expectations we do some

247
00:38:03.940 --> 00:38:15.310
which is called the naive machine learning estimators. We do some parametric adjustment. In particular, we use the doubly robust estimator.

248
00:38:16.760 --> 00:38:18.719
Okay, so

249
00:38:19.590 --> 00:38:30.400
here's the result. And let's jump to the highlights of the result. And the first. if we see the contrast.

250
00:38:30.700 --> 00:38:36.200
the contrast between unadjusted and majestic gap, you can see that.

251
00:38:36.230 --> 00:38:43.300
 so you can see a larger change.

252
00:38:43.640 --> 00:38:50.640
larger change for adjusted gap at the first period, and the smaller change in the

253
00:38:50.680 --> 00:39:17.469
 smaller changing the adjusted gap in the second period. And this is because that during the 1970 s. The adjusted gap is larger than the unadjusted gap. This means the negative selection of the workforce of female workforce, or a severe discrimination, and during the 1990 s. The

254
00:39:17.580 --> 00:39:28.599
estimation of unadjusted and adjusted gap are just similar, and again, a just a gap is larger. In recent years, which means

255
00:39:28.780 --> 00:39:31.690
the the trend

256
00:39:32.030 --> 00:39:33.809
by adjusted gap.

257
00:39:34.190 --> 00:39:45.210
Start in recent years, but for unadjusted gaps that we do not, if we do not control for human capital. It continues to converge.

258
00:39:46.330 --> 00:39:52.850
Okay? And then that's focusing on the result for adjusted gap.

259
00:39:54.290 --> 00:39:59.479
We have 2 sets of comparison. First, if we look at the

260
00:39:59.750 --> 00:40:08.390
 comparison between within occupation, gender, wage gap, and the change in the

261
00:40:08.550 --> 00:40:10.710
occupational segregation.

262
00:40:10.750 --> 00:40:13.919
which is the second and the third bar.

263
00:40:16.090 --> 00:40:28.999
What we. What I want to highlight is that changing within occupation, gender wage gap is always the most important equalizer. across 2 time periods, and

264
00:40:29.110 --> 00:40:34.399
the change in occupational gender segregation. It's

265
00:40:34.430 --> 00:40:36.349
the equalizing role in

266
00:40:36.450 --> 00:40:57.170
closing general which gap it is less significant than the general general impression. However, the stalling out of this aggregation process has been consequential, as you can see that in the from the period 1990 s. To 2,010, it becomes

267
00:40:59.300 --> 00:41:09.409
okay. I think I should. It becomes a T-s equalizer. Not a lot. But

268
00:41:09.640 --> 00:41:11.460
it's very trivial.

269
00:41:11.690 --> 00:41:13.919
at this period.

270
00:41:16.440 --> 00:41:25.089
And finally, if we look at the change in the wage structure and composition, and we can see that

271
00:41:25.720 --> 00:41:29.780
collectively they are trivial quantitatively.

272
00:41:30.290 --> 00:41:56.239
and the implication is important for this seemingly trivial number, that is this offers a bottom line estimate that the emerging literature which hypotheses that skill, bias, technological change, and the industrialization is family female, friendly. Well, they found the partial effect. However, we found that that bottom line is the overall effect is very small.

273
00:41:56.650 --> 00:42:08.639
Okay, we also do the counterfactual simulation and robustness checks. I'll leave this to Q. And A. Since we only have a couple of

274
00:42:08.680 --> 00:42:10.420
a few presenters.

275
00:42:10.840 --> 00:42:14.110
okay, thank you. I'll stop here.

276
00:42:15.020 --> 00:42:15.900
You very much.

277
00:42:20.370 --> 00:42:22.539
you know question of the light at all.

278
00:42:29.050 --> 00:42:46.140
So the last paper going to be a pre-recorded one. The the papers title is Gender Gap Dynamics among Refugees, and the recent immigrant immigrants. Different start different

279
00:42:46.250 --> 00:42:47.480
patterns.

280
00:42:47.590 --> 00:42:51.309
So this can be presented by Yulia.

281
00:42:51.440 --> 00:42:58.450
Because, yeah, Kova, she's from Herbold University Building, Germany.

282
00:43:14.180 --> 00:43:16.610
Sorry. Usually when I click on them. They just start playing.

283
00:43:26.090 --> 00:43:43.960
Hello, everybody! My name is Yulia Kosakova, and I will present you today a project entitled Gender Gap Dynamics among the refugees and recent immigrants, different styles, similar parties. And this is a joint work with still in the liquid troop from Mood university building.

284
00:43:44.970 --> 00:44:00.639
Okay, let's start with Defining what is what we know about female refugee gap. In the literature there is pronounced labour market disadvantages among refugees in various countries compared to other immigrants.

285
00:44:00.830 --> 00:44:18.249
and particularly integration of female immigrants, are behind their male counterparts, and that is moved even more for female refugees. This has been also attributed, or termed a so-called triple disadvantage for women who arrived as refugees.

286
00:44:18.290 --> 00:44:29.019
because, when competing for jobs, female refugees face disadvantages attributed to gender immigrant stages and force migration simultaneously

287
00:44:29.340 --> 00:44:45.109
and up-to-date. Little is known whether and to which extent the refugee and other immigrant women face district challenges in labour market integration, and how the employment. Gender gap develops over duration of state.

288
00:44:45.250 --> 00:45:12.230
I will focus on Germany. So migration and immigration in Germany is driven by 2 particular groups. One is migration from you 28 countries, particularly from new Member States and Eastern Europe. So between 2,010 and 2,020 almost 2.4 million people arrived, migrated to Germany.

289
00:45:12.480 --> 00:45:25.780
One certain of them are women, and they possess, on average high levels of education and occupational qualifications. What is important. You 28 immigrants have unrestricted legal market access.

290
00:45:25.970 --> 00:45:40.979
The other important source of migration are non-new nationals, which peaked in 2,015 to 2,016, due to 1.2 million first asylum applications submitted in Germany.

291
00:45:41.400 --> 00:46:00.509
there are no new nationals. consists from to one third of refugees, then, roughly, 15% are immigrants for family reasons. 10% arrive for education and training purposes, and only 8% arrived for employment purposes

292
00:46:00.960 --> 00:46:21.399
among non-new nationals. Roughly, 42% are women, and generally the I don't know. Nationals have non-restricted labour market access among economic and family migrants, for there is also under secret labour market access for proved refugees, and there are some restrictions for those who you arrive for education purposes.

293
00:46:21.670 --> 00:46:50.539
However, there are a lot of prolific obstacles in excess for refugees with UN clarified and tolerated status in Germany, and based on these 2 sources, the decision was not to distinguish between you and don your Nationals, but because of different labour, market access and different restrictions. we decided to distinguish between refugee and non-refugee immigrants.

294
00:46:50.700 --> 00:47:08.290
So it will look on the official data, we observed that basically the employment rates of non refugee migrants here are other foreigners. They are on a high level compared to

295
00:47:08.560 --> 00:47:25.709
so-called foreigners from war and crisis countries. and to we observe that there is a for among normal refugee immigrants there is some pronounced small pronounce gender gap, which is constant over time.

296
00:47:25.880 --> 00:47:46.659
When we look on the employment rates of of refugee immigrants, or those who arrived from war and crisis countries, we observed that The employment rates went back in the mid of 2,015, particularly driven by the high influence of refugees from these countries, and then

297
00:47:46.660 --> 00:48:12.189
we observed that man improved over time a lot achieving in 2,019 the level of non refugee women, while refugee women continue to perform on a lower. There are various explanation explanations for greater disadvantage of immigrant women and refugees, and also generally

298
00:48:12.190 --> 00:48:26.909
women. I will not go through them in detail, but generally, we can distinguish between explanations based on human capital and transferability issues. So here, also important to mention that

299
00:48:27.340 --> 00:48:45.979
women make gender-specific human capital investments in anticipation of family and key obligation which affect not only level of their education, but but also type of human capital. They bring when they arrive into destination country.

300
00:48:46.030 --> 00:48:58.319
Then there are issues or explanations based on self-selection, on education here, particularly that immigrants are on average or non-refugee. Migrants are on average.

301
00:48:58.440 --> 00:49:11.460
better self-selected, based on education compared to refugee, to refugees. and then other important aspects are language skills before and after migration. Here also.

302
00:49:11.860 --> 00:49:30.230
research shows that refugees generally arrive without any language skills of destination country which is mainly driven by their unprepared unpreparation of migration or no possibility to to prepare migration.

303
00:49:30.780 --> 00:49:52.429
Further explanation referred to family roles and care responsibilities particularly so. These arguments are more important for women. and here the series states that family roles and care, responsibilities and values are driven by the culture of the oraging country.

304
00:49:52.480 --> 00:50:03.360
and that here also important, like traditional roles, child care and villain is to be active on the labour market is on average, it should be low among refugee women.

305
00:50:04.180 --> 00:50:20.629
First explanation relate to social network and embeddedness on the import of of social contexts. With natives. health is less understood in this regard, but it becomes more and more important factor that

306
00:50:21.200 --> 00:50:50.289
provides important explanation for differences in employment rates among these different groups, and the final, which is particularly relevant for for refugees, relates to legal and labour market constraints, and here, in particularly at the lens of asylum procedure, during which refugees often have no access, not that labor market, no language courses

307
00:50:50.420 --> 00:51:22.079
but also important aspects, relate to context of reception. but hostility and prejudice of the majority of population. So to understand and to examine the gender gap among refugee and non-refugee immigrants will relate on 2 highly comparable data sources. Both are longitudinal data sources. So for refugees who rely on the ebay bomb sub refugee sample of refugees in Germany.

308
00:51:22.190 --> 00:51:34.349
for years 2,016 to 2,019, and for other immigrants who rely on yeah. So of migration sample lynched, launched in 2,013.

309
00:51:35.340 --> 00:51:46.670
For both in datasets we restricted samples to individuals in working age with maximum of 6 years in Germany at the time of their first interview.

310
00:51:47.070 --> 00:52:02.729
and based on these restrictions for packages we have slightly more than 7 so 1,000 individuals, and for other immigrants we have slightly less than 2,000 individuals.

311
00:52:04.360 --> 00:52:18.979
our dependent on arrival is gain for employment. Here we consider full of part time employed in vocational education or internship, or a presentership, and those marginally employed.

312
00:52:18.980 --> 00:52:40.970
If growth, spons, and earnings are above 0. We have based on the provided explanation before we have a lot of explanatory constructs. So we measure demographics. Human capital. Here is we look on the pre-migration, post-migration, human capital attainment.

313
00:52:41.030 --> 00:53:03.370
They were market experience in the country of origin, German language skills before and after migration. Then we have various variables for care, responsibilities. We also have values and motivations, social contacts, health status, legal stages, and also some regional and data controls.

314
00:53:04.760 --> 00:53:20.960
We rely on the Hacker line, the the composition method of the employment Gender Gap. And here the explained part relates to the part of the employment and the gap due to compositional differences in the characteristics from estimates.

315
00:53:21.000 --> 00:53:34.810
And here assumption is that the same assumption is same, returns to characteristics or endowments of men and women. The unexplained part of the employment gender gap relates or

316
00:53:34.860 --> 00:53:52.370
arises, due to the different effects of the same characteristics or endowments. In those groups we have 3 analytical steps. First, we look on, explained and unexplained portions of the employment gender gap for refugees and other migrants

317
00:53:52.380 --> 00:54:03.560
separately. then we look on the development of the employment gender gap, or by duration of stay in Germany.

318
00:54:03.990 --> 00:54:24.259
And then we look on the contribution of different constructs to employment, to explain employment, gender gap among those groups is that women, refugee women in particular have challenge in starting possession, so compared to male refugees, female and male.

319
00:54:24.350 --> 00:54:47.980
Among other immigrants, they possess lower educational attainment. On average they arrive with different types of human capital, particularly they were more likely to be employed in education sector before there are migration which is likely to be difficult and transfer to regulated labour market. In Germany.

320
00:54:48.860 --> 00:55:07.360
refugee women have also lower German skills, small care, responsibilities, fewer contacts beyond family networks and deal with more health issues. This is the first analysis, descriptive analysis on the employment gender gap, or the duration of stay among refugee and non refugee

321
00:55:07.360 --> 00:55:20.090
immigrants, and what we are, what we observe is striking similarity to their official registered data, which I showed you in the vigor a couple of minutes before.

322
00:55:20.760 --> 00:55:43.369
So for so first we observed that let's say, refugees generally are less equipped with labor market relevant resources at rival compared to other immigrants, and this is reflected in refugees. Lower employment rates at arrival in Germany, particularly for refugee women.

323
00:55:44.310 --> 00:56:12.700
And then we observed that the refugee men improve a lot over time. While this is not the case for refugee women, so the gender, the initial law agenda gap is increasing. Refugee migrants will observe quite opposite trend here initial. A bigger gender gap somewhat reduces over time.

324
00:56:12.860 --> 00:56:29.840
So we turn now to the results from the Ahhaka blind. The decomposition, and we observed that quite similar. There is a 2627 percentage point employment gender gap among refugees and other immigrants.

325
00:56:30.110 --> 00:56:42.970
and it will look on the explained part. we observed that the model is more suitable to explain the refugee gender gap than it is the case for other inference.

326
00:56:44.420 --> 00:57:02.990
when we turn to the dynamics or development of the gender gap over duration of state, we observe that among refugees the employment gender gap increases while the share of explained path reduces

327
00:57:03.670 --> 00:57:38.530
among non refugee immigrants. We've observed the opposite, that the gap in employment and the gap among them all, or reduces over time, but also the share of the explained parts. I'm looking on the components of the employment gender Gap among both group. I would provide you several important passages. So the important drivers of the employment Gender Gap are the pre migration, labour, market experience of different gender differences in prey, migration.

328
00:57:38.530 --> 00:58:02.579
labour, market experience, particularly at the initial stages, and this is particularly the case of refugees. which is mainly driven by the fact that women were working before migration in sectors which are slightly, really difficult and challenging to transform in the German labour market.

329
00:58:02.740 --> 00:58:12.830
Also, language skills and social contacts seems to be important drivers of the employment Gender Gap and its development.

330
00:58:13.100 --> 00:58:33.349
and particularly among refugees, health is important factor for both, and care responsibilities. What we observed is a challenge in starting position of refugee women, due to their lower endowment, with resources necessarily to gain to a foot halt in the labour market.

331
00:58:33.430 --> 00:58:54.010
and we also observed a growing employment gender gap among refugees, which was mainly driven by improvements of male refugees, and we also observed somewhat reducing gender gap. Among other immigrants, which was driven by improvements of female and refugees.

332
00:58:54.850 --> 00:59:08.999
and basically rather than the level of educational attainment. It is a labor market experience and the type of job before migration, which seems to play a role in the development of the agenda gap among refugees.

333
00:59:09.230 --> 00:59:28.569
How, if ever, in the middle and long run, it is language, proficiency, carrying responsibilities, social contacts outside their family as well as physical and mental health deficiencies that seem to be responsible for female refugees consistently lower employment rates.

334
00:59:29.320 --> 00:59:48.269
However, what we also like to stress is that the proposed theoretical constraint were unable to solve the puzzle. for observed, pronounced residual employment gender gap which even increases over duration of state, particularly for refugees.

335
00:59:48.330 --> 01:00:03.809
and this could reflect either on observed motivation and preferences or discrimination which we cannot. They've got with our analysis, but we cannot also exclude this as an explanation.

336
01:00:03.820 --> 01:00:13.870
and what is more important is that on observed resources of preferences or discrimination seems to become even more pronounced after arrival.

337
01:00:14.640 --> 01:00:22.090
Thank you very much. And I'm happy to have your questions. So these are contacts. How you can reach me.

338
01:00:26.070 --> 01:00:38.529
Okay, thank you very much. I wish you, we can, you know, ask a questions to her, you really as well, but unfortunately, we have to an impression presentation. I'd like to invite to the joy. And

339
01:00:38.710 --> 01:00:47.660
she come here. And you know I like to open the floor asking questions and then answering.

340
01:00:57.000 --> 01:01:02.200
okay, so feel free to ask questions. And the first I'd like you to.

341
01:01:02.220 --> 01:01:05.090
you know, through your names. And that's the

342
01:01:05.200 --> 01:01:05.890
this.

343
01:01:07.990 --> 01:01:08.720
Yeah.

344
01:01:15.870 --> 01:01:36.870
thank you very much for your presentations. I have a question on G. I think that your your paper is a great contribution to the research on the stalled Gender Gap, which is a puzzle for everyone. Why doesn't there some some change going on after this long term of revolution? I have a question more on the technical part.

345
01:01:36.870 --> 01:01:46.990
You decided to look at full-time career wages only what happens regarding part-time employment. So

346
01:01:47.300 --> 01:01:56.160
what about women working in part-time? Is it an important group that you like, or is it not the case in the period you're observing the what about part-time women?

347
01:01:56.390 --> 01:01:57.390
Oh, well.

348
01:01:57.850 --> 01:02:00.469
actually, actually.

349
01:02:00.750 --> 01:02:03.650
if we include a part-time?

350
01:02:06.110 --> 01:02:07.160
Okay, thank you.

351
01:02:07.210 --> 01:02:09.190
So the overall wage

352
01:02:09.270 --> 01:02:18.550
overall role. Unadjusted wage gap trend is very similar, for if we include the

353
01:02:18.690 --> 01:02:40.339
part-time workers, but we do not include them in the in the detailed analyzes because we want to have a conversation make our results more comparable with previous research. And they generally focus just so focus on the population of full time four-year workers.

354
01:02:40.380 --> 01:02:55.290
Actually, this is the the gender wage gap is even larger among full-time four-year workers. So we can also substantively. We can also like.

355
01:02:55.550 --> 01:03:01.620
focus more on the problem. Where's where the problem is more severe?

356
01:03:08.970 --> 01:03:10.630
Thank you for a question.

357
01:03:19.940 --> 01:03:25.640
Hi, my name is Pigeon Cole from Monash University. I have a question for you.

358
01:03:25.770 --> 01:03:49.980
I say, I miss the first part of the present. Post-power presentation, but I really like your paper. To tackle benefit receipt. So my question is more theoretical. so maybe I first I miss the first part. That's why I didn't understand. first of all, the social benefits that you conceptualize seems to be a short-term receipt.

359
01:03:50.010 --> 01:04:08.749
because I was also reading your paper like. So I wasn't very sure about if there is a duration is a process of receiving the benefits, because sometimes for people who are in need of, we see maybe just a short half a year lengths over.

360
01:04:08.810 --> 01:04:23.410
Would you come that as like this group of people of us, having, like we see, the social benefits for a long time, or just, or do you just group them all together? I think it would represent 2 different policy implications.

361
01:04:23.490 --> 01:04:48.019
And the second thing is that I was also looking at the migration backgrounds like how you construct those different migration groups. Then I read the sentences that you say that if in a family in a Dutch family. If you basically put some mother's births, the first country to quote whether this person will be the second generation of

362
01:04:48.380 --> 01:05:02.810
this person would be the second generation migrants, and if that mother is a Dutch, they will you. You focus on father's first country. So I'm just wondering whether if you consider that there might be a

363
01:05:02.960 --> 01:05:05.419
if we are talking about a simulation series.

364
01:05:05.810 --> 01:05:19.620
especially your father actually is native actually, already. It's an incident advantage to the to the Migration family. So I'm just wondering how you think about that.

365
01:05:19.820 --> 01:05:31.909
because for now, in your analysis they seem to treat as soon as there is a parent who belongs to non-douge family. That person belong to the second generation of migrant families.

366
01:05:31.950 --> 01:05:34.799
But I think theoretically there might be some people.

367
01:05:34.980 --> 01:05:35.820
A

368
01:05:40.870 --> 01:05:41.670
was fun.

369
01:05:43.060 --> 01:05:50.609
Thank you so much for your questions. Let me first start with the trusting your questions about the social welfare.

370
01:05:50.680 --> 01:05:59.490
So I study social systems benefit, which is basically a benefit that you can receive unlimitedly.

371
01:05:59.540 --> 01:06:03.379
But the 10 is means tested. So it depends on whether.

372
01:06:03.410 --> 01:06:15.379
 how much wealth you have, for instance, to be able to apply for this benefit. The unemployment insurance is based on how many months you've worked before.

373
01:06:15.450 --> 01:06:20.399
So that's a need, a shorter duration than it can range up until 2 years.

374
01:06:20.440 --> 01:06:21.550
Benefit received.

375
01:06:22.460 --> 01:06:26.339
I currently chose one month, a year.

376
01:06:26.390 --> 01:06:28.160
and as a minimal

377
01:06:28.500 --> 01:06:34.289
indication of benefit, you see. But we do intend to also run models with a long regulation of any

378
01:06:34.790 --> 01:06:46.040
 that answers your first question, then your second question with regard to migration background. Indeed, I used a more classical way of personalizing second-generation migrants

379
01:06:46.130 --> 01:06:56.279
and thinking about it. It might actually be very interesting to model the origin of the parents and mother, and so simultaneously in the model

380
01:06:56.450 --> 01:07:01.340
which these mighty downloads, these which I presented today, would be very flexible.

381
01:07:01.850 --> 01:07:06.870
And so it might actually be a very interesting avenue for future research.

382
01:07:07.290 --> 01:07:09.300
thinking about it conceptually.

383
01:07:09.700 --> 01:07:16.650
 whether or not actually the father's background is more important.

384
01:07:16.780 --> 01:07:18.830
4 children

385
01:07:22.680 --> 01:07:37.109
like it also have to do also to what extent its children can acquire social resources in the Netherlands. and I think both Monos and partners play a very important role with regards to

386
01:07:37.150 --> 01:07:39.879
How do you help with

387
01:07:40.040 --> 01:07:46.500
education? It might be that when it comes to to wealth and income, and there's a single

388
01:07:46.510 --> 01:07:58.650
breath in a system in the family, then, indeed, the father's employment and income and wealth might be very important here.  to come back to your question. I think it's a very interesting potential avenue for future research.

389
01:08:06.690 --> 01:08:20.249
Good morning, everyone. This is, too, for me. Patrick Donald from me here. Thanks for 2 interesting presentations. I actually want to ask. gee, I a follow up question to what Murray to alcohol mentioned earlier.

390
01:08:20.460 --> 01:08:38.119
I'm a bit worried if you exclude part time work altogether in this in this kind of analysis, because if we think about compensating differentials, explanations for the gender wage gap and women select into occupations where they're more likely to work full time.

391
01:08:38.480 --> 01:08:46.950
I'm sorry. Part time, then full time. I I fear that you might be missing a lot of explanation for the wage gap. So

392
01:08:46.990 --> 01:09:03.390
my question is, have you so? Do you have a better justification for selecting only full-time women than just? We want to follow the literature that already exists? Because if you don't, I really think you should be including part-time work, or I feel you're really missing a very large

393
01:09:03.510 --> 01:09:05.469
part of the the puzzle.

394
01:09:07.270 --> 01:09:20.239
Yeah. thanks for your I think you have offered a suggestion. I will do a so more additional sensitive sensitivity analysis to include a sample of

395
01:09:20.350 --> 01:09:23.720
but part-time part-time workers or control for that

396
01:09:23.930 --> 01:09:25.170
But

397
01:09:25.190 --> 01:09:31.750
from the conceptual perspective, I think we are actually looking at a population of

398
01:09:31.990 --> 01:09:34.500
working groups. That's

399
01:09:34.970 --> 01:09:48.880
If you adjust it for selection like, do it with hackman, two-step adjustment. You are actually talking about problem with different population.

400
01:09:50.210 --> 01:09:55.909
 I I don't know if this makes sense. But

401
01:09:56.170 --> 01:10:02.020
yeah, I will. I will do. I will check for that part as well.

402
01:10:03.880 --> 01:10:04.980
Yeah.

403
01:10:12.190 --> 01:10:13.090
A,

404
01:10:14.050 --> 01:10:27.430
okay. And I need a question about your figure. So it suggests that. So I did not completely understand your analysis, but the figure suggested that you have an instrumental variable.

405
01:10:28.360 --> 01:10:36.580
At least it's at Iv below occupational competition. the figure that you had in the very beginning.

406
01:10:36.960 --> 01:10:38.470
So

407
01:10:38.490 --> 01:10:42.969
what was that? Do you have an instrument of their approach? No.

408
01:10:43.220 --> 01:10:47.199
okay. So what is, I mean for them.

409
01:10:48.680 --> 01:10:56.650
You mean the bigger of a conceptual model.

410
01:11:39.640 --> 01:11:42.220
Not that into identification.

411
01:11:45.710 --> 01:11:46.850
The

412
01:11:48.870 --> 01:12:12.030
I'm lind you from the tennis University of Hong Kong. So. this is our goals of paper. So the method because you are confused about the methods. So it's definitely not about instrumental work, but nothing about it, but rather what we are trying to do is to try to provide an exhausted and comprehensive decomposition of different type of occupation-related forces to differentiate them

413
01:12:12.230 --> 01:12:30.310
can separate their contribution apart, and to evaluate their relative importance to the change of generator. Get trend over the past half century, and that's what we are trying to do. And the decomposition method is actually not just developers. We actually view on previous work by

414
01:12:30.860 --> 01:12:42.749
the Broom and Joe Sean's work recently published in American Sociological Review. but they only have 3 components. Now we have 4 components. And that's what we have added to that literature.

415
01:13:05.510 --> 01:13:11.240
I understand that the change in gender ratio has been studied.

416
01:13:11.700 --> 01:13:12.670
This

417
01:13:12.950 --> 01:13:25.940
and the use of very different skew-based cell phones, or skewed by a State force, will change. But I like the way that you, for example, which is to take occupation very seriously.

418
01:13:26.030 --> 01:13:30.610
Cameras don't take occupations here. Sociologists do it.

419
01:13:34.800 --> 01:13:42.289
and I think you did a very good job. My question is, I think you mentioned at the very end, very briefly.

420
01:13:42.320 --> 01:13:53.220
what are the benefits of this occupation-based approach in? How do the findings complement the findings by economists?

421
01:13:54.450 --> 01:14:09.850
thank you. This is exactly one of our main contribution that we want to make that because economists dozen economists don't take occupation seriously. They

422
01:14:09.860 --> 01:14:11.409
tend to provide

423
01:14:11.420 --> 01:14:17.199
various mechanisms, but whenever it comes to occupation they just

424
01:14:17.790 --> 01:14:34.499
don't pay attention. Put effort to make clear cut evaluation. And as sociologists, we really want to know the relative importance of these different roles played by occupation, especially like

425
01:14:34.560 --> 01:14:56.180
we have long discussed the importance of occupational gender segregation, and we all thought it must be because the trend of occupational segregation just parallel the trend of genderwich G. We thought well, since it parallels. It must be, almost explain all of the change. But

426
01:14:56.200 --> 01:14:57.350
we

427
01:14:57.820 --> 01:15:09.160
but actually, there's very few serious effort to have a clear evaluation in contrast with other occupation-related forces.

428
01:15:09.220 --> 01:15:19.440
So this is what we complement the existing literature. And also also it's also economists, economists.

429
01:15:19.630 --> 01:15:22.249
a. you know.

430
01:15:22.270 --> 01:15:31.010
the skill bias technological change. Literature is very influential across many subfields

431
01:15:31.130 --> 01:15:39.149
in labor economics, and in recent years some economists also

432
01:15:39.260 --> 01:15:44.899
in some research that they propose a hypothesis that this change

433
01:15:44.970 --> 01:15:55.249
female friendly, because the both the industrialization process and the skill-based Sbtc

434
01:15:55.280 --> 01:16:03.309
all contributed to the rise of service jobs and the decline in those manufactured jobs

435
01:16:03.370 --> 01:16:12.090
and manufacture jobs are dominated by male workers and service drops. There are more female workers.

436
01:16:12.130 --> 01:16:22.690
So this change, they, they look at those occupations and found that well, this seemingly gender-neutral impact

437
01:16:22.790 --> 01:16:30.339
has a female friendly outcome in terms of wage, but also

438
01:16:30.630 --> 01:16:35.739
but they didn't look at. They didn't take into consideration that

439
01:16:36.220 --> 01:16:39.289
the technological change also affect

440
01:16:39.400 --> 01:16:50.159
non-manual routine jobs. Like many, many kind of white-collar jobs, aside from service drops and

441
01:16:50.200 --> 01:16:58.009
the decline in those occupation, the decline of those occupations might affect.

442
01:16:58.130 --> 01:17:01.770
female workers are negatively.

443
01:17:02.110 --> 01:17:14.690
And so in this study we offer. we offer the bottom line for that that we, the composition, the component of changing composition and the component of

444
01:17:14.860 --> 01:17:20.410
changing occupational wage structure is the overall estimate of the

445
01:17:20.550 --> 01:17:22.019
outcome of

446
01:17:22.110 --> 01:17:30.129
this whole technological impact and the industrialization structural change.

447
01:17:30.830 --> 01:17:40.860
and we found that although the hypothesis is intriguing. the overall potential, the overall impact is very small.

448
01:17:42.110 --> 01:17:53.590
Thank you very much. I really like this kind of a challenge for the great to economists. Sometimes we can work together. But we we need a more, you know, strong division, or but anyway.

449
01:17:53.610 --> 01:17:58.490
and for the future. We we should. Oh, we can't collaborate and working together.

450
01:17:58.860 --> 01:18:00.520
But anyone.

451
01:18:03.100 --> 01:18:04.630
okay, one more.

452
01:18:07.750 --> 01:18:23.969
I have a rather simple question to the first presentation. I really like the idea of trying to test the hypothesis of intersectionality and then using a very articulate.

453
01:18:24.350 --> 01:18:34.930
I've seen a lot of papers on intersectionality on qualitative work. But this is very quantitative, and that's the part I really like.

454
01:18:35.950 --> 01:18:40.930
I haven't used those domains that you use. So I have a very naive question.

455
01:18:41.210 --> 01:19:00.459
The different combination of these these variables that I mentioned up in intersectionality as intersectionality. And then my question is, did you include all the possible interactions within the model or in the end, do you select just

456
01:19:00.730 --> 01:19:14.179
so significant combinations? We show the graphs? No, I think, do they include all the possible combinations, or simply just ones that you found with significant

457
01:19:17.940 --> 01:19:18.690
me.

458
01:19:18.990 --> 01:19:29.350
I first thank you for the compliment about starting now contain interception in church. It's not very common yet. But

459
01:19:29.460 --> 01:19:35.079
there are increasingly more number of papers also in sociology that study insectionally.

460
01:19:35.240 --> 01:19:43.970
 in my models I include all interactions and also the most significant ones.

461
01:19:44.250 --> 01:19:52.789
That's What is all the methodology that I have been using, and mainly is.

462
01:19:53.060 --> 01:19:56.729
is doing also just to account for

463
01:19:56.770 --> 01:19:57.870
and

464
01:19:58.750 --> 01:20:03.499
just have the most complete picture. So basically, I develop only an

465
01:20:03.670 --> 01:20:06.400
that is based on additive grounds, only

466
01:20:06.680 --> 01:20:09.709
modeling the gender immigration background in each effects

467
01:20:09.720 --> 01:20:15.069
in addition to each other, and then I have them all the way. I have all the interactions together.

468
01:20:15.260 --> 01:20:18.520
and then they compare those 2 estimates with each other

469
01:20:18.570 --> 01:20:19.760
and

470
01:20:20.010 --> 01:20:23.059
the model that you are. A

471
01:20:23.620 --> 01:20:25.669
proposing might also be available

472
01:20:26.070 --> 01:20:27.290
and

473
01:20:27.310 --> 01:20:33.870
potential method for quantitative intersectional studies to use machine learning or

474
01:20:34.190 --> 01:20:39.400
just ordinary. And these questions to see

475
01:20:39.470 --> 01:20:45.740
which interactions between those dimensions actually increase or decrease

476
01:20:46.160 --> 01:20:51.019
and the levels of inequality. In my case, for instance, benefit received.

477
01:20:53.100 --> 01:20:53.800
yeah.

478
01:20:54.430 --> 01:20:57.170
And there are many to that

479
01:20:57.260 --> 01:21:11.190
small thing, because I think maybe you can explain that much better than I do is that you do not interpret the actions themselves, but only the deviation from the

480
01:21:31.090 --> 01:21:35.089
group price like, in this case, 40

481
01:21:35.390 --> 01:21:50.220
different dummy variables for all the Unique donations of gender immigration account. And H in this model? and that's also what I really like about this methodology. You need to model all the interactions which would be computationally, also very and expensive.

482
01:21:50.790 --> 01:21:58.699
Imagine running a similar model with 5 or 6 dimensions and potential inequalities. And then you can also

483
01:21:58.720 --> 01:21:59.960
summarize the

484
01:22:00.320 --> 01:22:03.639
these interactions in just one term per intersection of straight and

485
01:22:03.820 --> 01:22:10.800
and this one term of the minimum additive effect, or the intersectional effect which I just used interchangeably.

486
01:22:10.850 --> 01:22:21.850
also very well captures the argumentation which have already been done in the qualitative field, so it not only has a computational gain, but it also is very much in line with the augmentation

487
01:22:21.870 --> 01:22:28.219
about disproportionate inequalities. The multiple jeopardy are leading to increased vulnerability.

488
01:22:28.300 --> 01:22:33.990
Then that intersectional effect components captures that increased vulnerability. Basically for that group

489
01:22:42.070 --> 01:22:43.160
there.

490
01:22:46.850 --> 01:23:05.299
I have a question actually, intersectionality, because it seems to me this concept is really, you know, popular. But at the same time we have done this kind of, you know, focusing on this kind of aspects, and using, you know, interaction or heterogeneity.

491
01:23:05.480 --> 01:23:07.850
But what's the most

492
01:23:07.950 --> 01:23:18.620
important or advantage to use the concept of intersectionality rather than interaction and heterogeneity. What do you think?

493
01:23:19.380 --> 01:23:28.230
 to the one end, I think, by using the term intersection, and it's also in a quantitative sense, we are bridging 2 fields together.

494
01:23:28.350 --> 01:23:35.189
and that they have been studying these inequalities. so that's a leak using interactions.

495
01:23:37.850 --> 01:23:51.589
And whereas the quality of scholars have been theorizing also about how these people might perceive their inequalities along the moment of the lines which have been overlooked. I would say, like on the day scholars for a long time

496
01:23:51.830 --> 01:23:52.880
and

497
01:23:54.000 --> 01:23:56.350
That will be my

498
01:23:56.830 --> 01:24:00.359
first answer to the question.

499
01:24:00.770 --> 01:24:05.760
And then also, I think molding interaction is not the right way to go for modeling.

500
01:24:05.770 --> 01:24:07.020
and

501
01:24:07.050 --> 01:24:10.290
it is actually inequalities per se, because

502
01:24:10.360 --> 01:24:15.920
and the qualitative field always have been very abundant about

503
01:24:16.040 --> 01:24:18.080
not using the white male

504
01:24:18.190 --> 01:24:20.859
old man as the reference category.

505
01:24:20.870 --> 01:24:30.889
And with the methodology that I'm using, you don't need to have a reference category because you're making within group comparisons based on this intersectional model. And this additive model.

506
01:24:31.090 --> 01:24:34.190
So it's a more flexible way of looking at these inequalities

507
01:24:34.230 --> 01:24:35.800
to summarize these

508
01:24:36.110 --> 01:24:40.810
potential momentative. That's in one component. And it makes work.

509
01:24:41.020 --> 01:24:45.129
The story that's more in line with the qualitative literature fuels.

510
01:24:45.320 --> 01:24:53.009
And it also very well relates basically the addition of these interactions back to the quantitative field

511
01:24:53.050 --> 01:24:58.990
which I've been studying already beyond. You have all these for general immigration up on H for a long time.

512
01:24:59.180 --> 01:25:03.060
and then you also see in one component. But then this approach actually adds to that.

513
01:25:04.440 --> 01:25:06.590
Very good. Thank you very much.

514
01:25:06.640 --> 01:25:12.620
Can I ask one more quick question that you're focusing on?

515
01:25:12.700 --> 01:25:14.810
We set up benefit

516
01:25:14.910 --> 01:25:19.080
not. You know whether you become an employer or not.

517
01:25:19.130 --> 01:25:27.039
Why step you? You've represented the receptor of the benefits that so? Why you choose this this response variable.

518
01:25:27.780 --> 01:25:36.729
 it's in one way, actually a convenience measure, because the research project is very much concerned.

519
01:25:36.810 --> 01:25:37.840
And

520
01:25:38.050 --> 01:25:47.239
the use of these provisions. like very social books in the Netherlands. Indeed, unemployment is a very important drawing of the

521
01:25:48.150 --> 01:25:57.320
 I think, actually, looking also at these benefit programs is very interesting

522
01:25:57.400 --> 01:26:07.890
because these programs are meant to basically and uphold the more economically vulnerable people to help them to sustain the way of living

523
01:26:08.240 --> 01:26:11.369
signs of unemployment. It actually also

524
01:26:12.940 --> 01:26:18.730
gives you some sense of like, how? How this

525
01:26:18.990 --> 01:26:22.410
support network is actually helping people out to more?

526
01:26:24.360 --> 01:26:25.470
Very good

527
01:26:25.570 --> 01:26:27.500
any other questions.

528
01:26:28.460 --> 01:26:42.079
I'm going to have a last question to Gee, actually, I really like your challenging, you know, and but at the same time again, why don't you take into account a part-time work as well, because

529
01:26:42.080 --> 01:26:58.699
again, they're looking at the waste differences. We can't ignore the increase in and part time work in the different situations, and you know the difference of the vanish and prominent status, or keeping in a proton, is also very much sociological point of view.

530
01:26:58.700 --> 01:27:10.530
Then again, you you show the change within within the occupational categories. It explains quite a bit.

531
01:27:10.570 --> 01:27:22.090
Then you conclude the main reason to reducing a wage gender wage gap, mainly because of the woman's increase of foot

532
01:27:22.410 --> 01:27:24.880
woman friendly, you know.

533
01:27:25.140 --> 01:27:40.889
No, no, this is officer. Okay? So I like to. You know, I like you to explain a little bit more on this mechanism, because I'm kind of confused because you're focusing on the female in a friendly occupation

534
01:27:40.940 --> 01:27:42.250
can help.

535
01:27:42.330 --> 01:27:47.470
you know, overall decrease in a wage gap or not.

536
01:27:50.380 --> 01:27:51.759
Thank you for a question.

537
01:27:51.800 --> 01:28:05.399
so I'll I'll manage your expectation at first that the the the main part of our analysis is offering a demographic accounting we cannot offer.

538
01:28:05.710 --> 01:28:10.060
Very sort of evidence that test the mechanism.

539
01:28:10.280 --> 01:28:11.230
And

540
01:28:11.480 --> 01:28:17.650
so what we are doing is actually answering some all questions and raising new questions.

541
01:28:17.730 --> 01:28:20.429
I should to the attention.

542
01:28:20.660 --> 01:28:23.289
 the the for the

543
01:28:23.710 --> 01:28:30.009
the part, you say the role of entering specific optical patients.

544
01:28:30.190 --> 01:28:31.810
actually.

545
01:28:32.190 --> 01:28:43.919
What we want to say is. What we found is that we used to. for for for scholars, those who thought that desegregation can solve

546
01:28:43.970 --> 01:28:51.679
everything, then it's not what we found that even with very rapid

547
01:28:52.050 --> 01:29:00.570
desegregation, that the the distribution of man women across different occupations are more equal, that

548
01:29:00.680 --> 01:29:07.089
the part that it translated into the closing Gen. Which gap is small.

549
01:29:07.370 --> 01:29:08.770
it's more and

550
01:29:09.100 --> 01:29:15.839
actually, what contributed more to the overall closing is the weeding part.

551
01:29:15.910 --> 01:29:27.129
For this part. For this part there are various mechanisms, and the one of famous one is apparently the

552
01:29:27.490 --> 01:29:34.679
many cold covert, not over covert discriminations within occupation.

553
01:29:34.830 --> 01:29:35.969
And the

554
01:29:36.040 --> 01:29:41.360
 yes, and for

555
01:29:43.010 --> 01:29:46.989
what we actually answered, here is that

556
01:29:47.960 --> 01:29:48.920
for

557
01:29:49.040 --> 01:29:52.720
during the period of stalling out

558
01:29:53.550 --> 01:29:56.860
we might expect we can.

559
01:29:57.040 --> 01:30:10.329
I'll find some bad guy that strongly, that that plays a strong, dis equalizing role. And actually that it's not the case.

560
01:30:10.540 --> 01:30:17.489
It's not the case. It's that. the the the true story is that the part

561
01:30:17.620 --> 01:30:25.829
such as the weeding occupation, gender, wage gap, the part that used to contribute greatly to closing the gap

562
01:30:25.910 --> 01:30:27.730
become also a

563
01:30:28.040 --> 01:30:29.410
so slow.

564
01:30:29.940 --> 01:30:42.139
So when we say we raise new questions. Is that that now we wonder what happened in this part, and

565
01:30:42.230 --> 01:30:57.919
the information we gave is that when we look at mechanism, our overall demographic accounting give you a map, a direction, that which one overall is more important and worth digging into more details.

566
01:30:58.160 --> 01:31:27.410
Very good, thank you very much. And then, lastly, I like to say that one more distributed papers, which is very interesting, and unfortunately, you know, there is no time to present, but I like to say. That last paper, you know, Will, would be the about the dash. Parental parental education, shape, gender wage gap over the life course, new insight from Australia.

567
01:31:27.410 --> 01:31:46.280
So I I really miss this paper. But you know, hopefully, you can. We all have a chance to read it? Okay. So if you do not have any kind of questions, a comment. I like to close it, and I let you go to the longer of lunch break.

568
01:31:46.370 --> 01:31:49.029
So no, thank you.

569
01:31:49.050 --> 01:31:52.000
Thank you very much.