Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=

# Masculinity Norms, Occupational Choice, and Reallocation: Micro and Macro Evidence

Authors: Rachel J. Schuh, Martina Uccioli  
Discussant: None  
Video: https://www.youtube.com/watch?v=1kb3a99sA-E&t=13246s


## Talk (03:40:46 – 03:57:01)

[03:40:46] finishing off our day with three egg timer presentations, 15 minutes each.  
[03:40:51] We're going to hold questions during the presentation, but fortunately, we have a dinner reception right after where you can ask the presenters all of your questions. So, I'm going to hand it over to Rachel who will tell us about masculinity norms. Please go ahead.  
[03:41:05] >> Okay. Thank you. So today I'll be presenting a brief overview of joint work with Martina Ucholi on masculinity norms, occupational choice and reallocation. And I'm going to start  
[03:41:17] with a fact on sectoral reallocation. If we look at employment growth by industry in the past 50 years or so, growth in historically more female jobs has vastly outpaced growth in historically more  
[03:41:31] male jobs. So this plot shows this by splitting industries by their female share in the late 1960s and then plotting cumulative employment growth in those industry groups. And we can see in the yellow and purple lines that  
[03:41:44] industries that started off more female have seen much more employment growth than industries that started off more male. Now at the same time as this growth in female jobs has occurred, male participation has declined dramatically.  
[03:41:58] So over the same time period, male labor force participation has declined from over 80% in the late 1960s to under 70% today. And this looks at aggregate male labor force participation. But the same  
[03:42:11] is true if we look at prime age men or if we look at the male employment to population ratio.  
[03:42:17] So the question that jumped out at us looking at these two trends was are they connected? Are some men leaving the labor force because opportunities are moving to female dominated sectors?  
[03:42:30] So we wanted to check just kind of a basic test. Are these two things correlated? And we find that they are when we look across locations and time periods. So this plot shows at the commuting zone decade level a measure of the growth of female dominated sectors  
[03:42:44] against changes in male employment to population ratios. And these are measured over 10-year periods from 1980 to 2010. And we can see that in commuting zones where female dominated sectors grew relatively more, male  
[03:42:57] employment fell relatively more. So then the question is why are these two things connected? Because it's not a priority obvious that they should be. It's possible that female sectors just grew because female labor supply increased  
[03:43:10] which wouldn't necessarily affect men. Or if even if demand is growing in female sectors, men could switch into those sectors. So this leads us to our proposed link between these trends which relates to gender norms. So here we just  
[03:43:25] split these communing zones up by whether they're in states with more or less progressive gender norms as measured in the GSS. And we see here that states with less progressive gender norms in red see a stronger connection between growth of female dominated  
[03:43:39] sectors and changes in male employment. So this leads us to our key question.  
[03:43:46] Why haven't men switched into growing female jobs? And a proposal for a missing link. So, I just showed three macro facts. First, female jobs are growing much more so than male jobs.  
[03:43:56] This is an ongoing trend, but it continues into the present day with the large growth in sectors like healthcare.  
[03:44:02] And at the same time, male employment and participation has declined. We find that these two trends are correlated and they're more correlated in more gender conservative locations.  
[03:44:13] So here we're going to propose a missing link, a reason why men haven't moved into these female dominated sectors, and that's masculinity norms and how they relate to occupation choice. So I'm just going to go into a brief aside before I  
[03:44:27] get into what exactly do what exactly I do. What are masculinity norms? So typically in research on gender norms and especially in largecale US data sets that have measures of gender norms, we look at things that measure appropriate  
[03:44:42] conduct for women. So we'll look for agreement with statements like a preschool child is likely to suffer if his or her mother works. So this is what's appropriate for women to do and specifically these questions often look at mothers.  
[03:44:55] So we're going to define masculinity norms as something slightly different following ongoing work work by Deas and co-authors as the informal rules that guide and constrain men's behaviors by shaping beliefs about appropriate conduct for men. So rather than what's  
[03:45:09] appropriate for women, we're looking at what's appropriate for men. Sometimes those will overlap directly and sometimes they'll be more separate. So, we'll measure agreement with statements like men should be aggressive and competitive to get ahead, but also with statements more directly relating to  
[03:45:24] work, like there is some work that is men's and some that is women's and they should not be doing each other's.  
[03:45:30] And we're particularly interested in how these masculinity norms might constrain men's job choices.  
[03:45:37] So specifically in this paper motivated by these macro facts, we'll ask whether masculinity norms act as a friction and occupation choice. And in the micro data, we'll attack this from two angles.  
[03:45:48] First, in observational data, we'll ask whether gender norms affect responses to a layoff. So when men are laid off, do their gender norms affect how quickly they return to employment and what sort of sectors and occupations they move  
[03:46:01] into? Then in ongoing work, which is currently in the pilot stage, so we're very very open to feedback on everything, we ask directly if masculinity norms affect occupation choice by eliciting these masculinity norms and having some information  
[03:46:16] treatments on them and also estimating the willingness to pay to be in different occupations.  
[03:46:22] So I'll first go into the observational evidence on response to layoffs. Here we'll be using the PSID and we use pretty standard layoff event studies here where we compare individuals that were laid off to individuals that were never laid off. We have individual year  
[03:46:36] and age fixed effects with no other controls and we use Callaway Santana estimators. I won't get into the details of the estimation because this is only a 15-minute talk. Um, but the key difference here is that we're going to estimate these separately on what we call gender traditional and gender  
[03:46:51] progressive men. And I wanted to get into a little bit of detail about how we measure these norms in the PSID because there are some important constraints.  
[03:46:58] First of all, the questions about norms are not going to be about masculinity norms. They'll be these more standard gender norms I talked about earlier that are typically questions about women's behavior. Second, in the PSID, these are in the caregiver supplement. So, they're typically asked to mothers. So, we'll  
[03:47:13] assign the norms of mothers to the fathers in the family. So, we very rarely have norm estimates for the man.  
[03:47:20] So relating this to our question on masculinity norms affecting men's behavior will rely on two assumptions.  
[03:47:26] One that these more general gender norms are correlated with masculinity norms which has been found in other research on masculinity norms and that there's assortative matching on these gender norms but it still will be a noisy  
[03:47:39] measure of um men's norms. Then we're going to look at two outcomes. One, are men employed or unemployed or out of the labor force following a layoff? Second, are they in at what we call a female industry or occupation? So these are  
[03:47:53] going to be industries and occupations above a certain threshold of female shares, which is the top tersile of female shares in 2005.  
[03:48:01] So first just looking at employment responses here, these look like pretty standard layoff event studies. The layoff occurs between year 0 and minus2, which is the point before it. So employment falls after a layoff as we  
[03:48:15] always see in these and then recovers gradually but doesn't ever reach its level before the initial layoff. And we see here when we split these by gender traditional men in red and gender progressive men in blue that these progressive men recover slightly faster.  
[03:48:29] They return to employment a little bit quicker after this layoff. And then these two kind of converge when we go a little bit further in time. And again, these are very noisy measure measures of norms and it's a fairly small sample once we limit to people who were laid  
[03:48:43] off and have a norms measure. So there's a lot of noise in these, not super significant, but we're also looking at the NLSY to kind of expand our analysis here. Then we want to understand if these men recover employment faster by  
[03:48:56] moving into these grow excuse me growing female dominated industries and occupations. So, first we just look at do you move into a female industry? And we see in blue that progressive men are more likely to be in  
[03:49:10] a female dominated industry after a layoff. So, it seems that when they lose their job, they're more likely to move into a female industry, but gender traditional men are no more likely to be in a female dominated industry after this layoff. And the same is true for  
[03:49:24] occupations. So this is we take this as suggestive evidence that these more progressive men are recovering their employment more quickly by moving into these growing female dominated jobs. But as I mentioned at the beginning, there's  
[03:49:37] some important caveats to this. One, these aren't masculinity norms. They're these broader measures of gender norms.  
[03:49:45] And two, we can't really get a causal effect of gender norms here. These might be associated with a lot of different things, which is why we move to our survey experiment where we're going to directly ask if masculinity norms affect the willingness to pay to be in a gender  
[03:49:59] conforming job. So, we have two versions of the survey.  
[03:50:03] One in the US and one in the UK. In the US, we're using Prolific. We're currently piloting it in the UK. This is going to be in the understanding understanding society survey and the innovation panel. So, this is sort of like the PSID in the UK. It's a  
[03:50:16] longunning panel, but they have a module re where researchers can submit questions. So, we'll actually be able to link people to a longer panel there.  
[03:50:25] [snorts] And we have three main objectives in this survey. First, we want to measure and potentially experimentally change masculinity norms.  
[03:50:33] To do this, we're going to elicit masculinity norms by asking for agreement with a number of statements on masculinity. Most of these are taken from prior research measuring these norms, but we're going to add some new statements that specifically relate to  
[03:50:46] job choice. Then we're also going to have an information treatment in an attempt to debias beliefs about other people's masculinity norms, relying on the idea that there's some pluralistic ignorance here. So, people aren't necessarily well-informed about other  
[03:51:01] people's agreement with these masculinity norms. And I'm going to present some pilot results here. This is from a pilot of about 200 people, so they're all quite noisy. So don't take them as final results just as suggestive of what we will be able to show and  
[03:51:15] hopefully a taste of future results in the next couple of months. So first looking at these masculinity statements these lines just show the resp percent of respondents that agree with each  
[03:51:29] statement. So things like it is important for a man to take risks or men should act strong even if they feel scared. And the important thing to note just on these averages is that for most of these statements a non-trivial percent of respondents are agreeing with  
[03:51:44] them. So these are giving us somewhat some sort of measure that differs across individuals of norms. It's not that no one ever agrees with these statements even though they are in a lot of cases fairly. One might say extreme. And if you include the false statement they're  
[03:51:58] a little bit more extreme at times. Now, when we ask people what percent of people of their gender and education level they expect to agree with these statements and for each respondent, we'll only ask them about one statement.  
[03:52:11] We don't do that across all statements. We see that these expected agreement levels are much higher. So, for example, if we look at it's important for a man to take risks, about 30% of people agree with that, but they expect that over 50%  
[03:52:25] of people agree with that statement. So this suggests to us that there is scope for a debiasing inter intervention.  
[03:52:32] There's also a wide dispersion in these expected percent agreements. So this tells us that there's also potentially potential for treatment effect heterogeneity here if we give people information on these true agreement  
[03:52:46] rates. Okay. So next in the survey, we want to compute the willingness to pay among men to avoid a gender conf non-conforming occupation. So here we're going to do a  
[03:52:59] hypothetical job choice exercise where people choose between different jobs where we hypothetically offer free training and a guaranteed job at a given salary in different gendered occupations. So we use prior research  
[03:53:13] and some side surveys to pick some occupations that vary in their gender p perception. So, do people see this as a feminine job or a masculine job? And then we add this training component because we know a lot of these people are already in a specific occupation.  
[03:53:27] So, just asking you to switch into a new one with no training would be somewhat unrealistic. So, again, I'll show some pirate pilot results. Very noisy, but suggestive. So, here I'm going to look at the willingness to pay to avoid a  
[03:53:40] feminine occupation relative to a genderneutral occupation among men. Here the feminine occupations are going to be nurse and teacher and the genderneutral occupation is loan officer. And one thing we're very open to suggestions on  
[03:53:54] is which occupations we actually do because which occupations we choose is going to be important for this. So men on average have a willingness to pay of about $7,000 to avoid these female occupations relative to this genderneutral occupation. So the wage in  
[03:54:08] the female occupations would have to be $7,000 higher for them to be indifferent. And if we split this by men who score lower and higher on this masculinity scale, we find that more masculine men have a higher willingness to pay to avoid these feminine  
[03:54:23] occupations. Right now, these are very noisy, but we take this as suggestive that there's potentially something going on here. So, one concern with this design is that people already have skill investments in given occupations. So we might be partially picking up on the  
[03:54:38] fact that men have already invested skills in certain types of jobs and these female jobs are more distant from their current jobs. So we attempt to separate that out by having another module where we tried to get respondents to separate their occupation preferences  
[03:54:52] from their existing human capital by having them rank recommended occupations for their child. So we think of this as a situation where their children or a hypothetical child that they may care about um is still in school, hasn't made  
[03:55:05] as many human capital investments yet. So we can maybe get at more of these norms rather than just do they already have human capital investments in these different occupations. So again, I'll show some very noisy pilot results here.  
[03:55:19] So here I'm going to show willingness to pay iss to avoid a female occupation, but I'm going to show them in percentage points of future growth instead of in wages um because our wage data is a little um noisy right now for these, but you can think of them as similar. If  
[03:55:33] it's positive, that means we want to avoid the female occupation. We would prefer the genderneutral occupation. So we can see that respondents on average are willing to pay more to avoid the feminine occupations for their sons.  
[03:55:45] Here the occupations are going to be nurse and speech therapist compared to doctor or lawyer. So we're willing to pay more to avoid them for sons than daughters. Interestingly, if we split this by women and men respondents, women  
[03:55:59] are actually willing to pay more to avoid these occupations for their sons than men are. But if we split up men by more and less masculine, again, the more masculine men are willing to pay more to avoid these feminine occupations for  
[03:56:12] their son. So in my remaining seconds I will wrap up. In this paper we ask if masculinity norms are a friction to occupation choice. We show in the macro data that there's been sectoral reallocation towards female jobs and decreasing male employment. These are  
[03:56:26] correlated more in more gender conservative areas. In the micro observational data after a layoff progressive men's employment recovers faster via switching to feminine industries and occupations. We also show in the paper that children of traditional mothers are more likely to  
[03:56:41] choose gender stereotypical jobs. And in the experiment, we hope to show eventually that stronger masculinity norms are associated with or cause in the more experimental setting some willingness to pay to avoid a feminine occupation for oneself and some  
[03:56:55] willingness to pay to avoid a feminine occupation for one's son. So, thank you.